Assess Now ESE A: sEF 2 E

||||||||III US005 68262A
United States Patent (19) 11 Patent Number: 5,168,262

Okayama 45 Date of Patent: Dec. 1, 1992
(54 FIRE ALARM SYSTEM 4,727,359 2/1988 Yuchi et al. .................... 340/870.17
4,818,994 4/1989 Orth et al...... ... 340/870.21
75 Inventor: Yoshiaki Okayama, Tokyo, Japan 4,881,060 1 1/1989 Keen et al. .......................... 340/51
73) Assignee: Nohmi Bosai Kabushiki Kaisha, FOREIGN PATENT DOCUMENTS
Tokyo, Japan
59-157789 9/1984 Japan.
21 Appl. No.: 543,851 59-172093 9/1984 Japan.
(22) PCT Filed: Dec. 1, 1989 61-98498 5/1986 Japan.
86 PCT No.: PCT/JP89/01210 Primary Examiner-Donnie L. Crosland
Attorney, Agent, or Firm-Wenderoth, Lind & Ponack
S371 Date: Jul.18, 1990 57 ABSTRACT
S 102(e) Date: Jul.18, 1990 A fire alarm system employs a neural network for ob
30 Foreign Application Priority Data taining one or more types of fire related information
Dec. 2, 1988 JP Japan ..... ... 63-30477 values. A plurality of detection information values are
Dec. 8, 1988 JP Japan ................................ 63-3O8807 time-serially collected from plural fire phenomenon
detectors. The detection information values are signal
51) Int. Cli.............................................. G08B 23/00 processed such that a weighting coefficient is assigned
52 U.S. C. .................................... 340/523; 340/505; thereto in accordance with a relative significance of the
340/510,340/511; 340/514; 340/588 detection information value to the desired fire related
58 Field of Search ............... 340/523, 506, 510,511, information value. The various weighting coefficients
340/.514, 870.16, 870.17, 870.09, 870.21, 505, are stored in advance in a memory. The weighting coef
825.08, 588, 589 ficients stored are established so that the fire related
56 References Cited information value for a particular set of detection infor
mation values approximates a desired fire related infor
U.S. PATENT DOCUMENTS mation value.
4,556,873 12/1985 Yamada et al. ..................... 340/51
4,668,939 5/1987 Kimura et al. ... ... 34.0/51 11 Claims, 13 Drawing Sheets
- as a re- - -a rere rere- a-- re- - -- a re-3- re-a- - - -
DE:
FRE DETECTOR
u
processor
sts
c2NE.
- PROCESSOR
WORKAREA PROG
PROGRAM ROMZEERAMa ROM Ele

lconSTANT AB
- C RAM W
ORKAREA
assess now ESE a

SENSOR PART
CWN
ROM12 e st
TERM AOD AB
RAM2 DEF AB
s EF 2 E. ROM13 E. QXRAM3
WEIGHT VALUE
DISPLAY
---------------.-
SG
TRANSCEIVER F 1 - EF 12 Ed P
L (EIF 13 K ky EACH-N
SIG TRANSCEIVER DAANPU
EN KEY
U.S. Patent Dec. 1, 1992 Sheet 2 of 13 5,168,262
FIG. 2
OME SLV SLW2 SLV3 SLV4 SLV5 SLV6
INPUT 0.070 0.070 0.070 0.070 0.070 0.070
OUTPUT (T) 0.000
OUTPUT (R) 0.00
INPUT 0.070 0.30 0.90 0.260 0.320 0.380
2 OUTPUT (T) 0.200
OUTPUT (R) 0.56
NPU 0.070 0.90 0.320 0.450 0.580 0.630
3 OUTPUT (T) 0.600
OUTPUT (R) 0.604
NPUT 0.070 0.260 0.450 0.510 0.580 0.630
4 OUTPUT (T) 0.600
OUTPUT (R) 0.598
INPUT 0.260 0.320 0.380 0.450 0.380 0.580
5 OUTPUT (T) O.800
OUTPUT (R)
NPUT 0.260 0.260 0.260 0.260 0.260
6 OUTPUT (T)
OUTPUT (R)
INPUT 0.90 0.30 0.30 O,070 0.070
7 OUTPUT (T)
OUTPUT (R)
INPUT 0.450 0.50 0.580 0.50 0.580
8 OUTPUT (T)
OUTPUT (R)
INPUT 0.380 0.380 0.380 0.380 0.380
9 OUTPUT (T)
OUTPUT (R)
NPUT 0.320 0.260 O. 90 0.30 0.070
10 OUTPUT (T)
OUTPUT (R)
NPUT 0.260 0.30 0.070 0.30 0.070
11 OUTPUT (T)
OUTPUT (R)
NPU 0.540 0.580 0.630 0.540 - 0.60
12 OUTPUT (T)
OUTPUT (R)
NPU 0.510 0.50 0.50 0.50 0.510
3 OUTPUT (T)
OUTPUT (R)
FIG. 2
PATTERN
NUMBER SLV SLV2 SLV3 SV4 SLV5 SLV6
NPUT 0.50 0.450 0.380 0.320 0.260 0.90
4 OUTPUT (T) 0.300
OUTPUT (R) 0.238
INPUT 0.50 0.380 O.260 0.30 0.260 0.90
5 OUTPUT (T) 0.00
OUTPUT (R) 0.037
INPUT 0.070 0.320 0.070 0.30 0.30 O. 90
16 OUTPUT (T) 0.000
OUTPUT (R) 0.006
NPUT 0.070 0.90 0.510 0.450 0.380 0.380
7 OUTPUT (T) 0.300
OUTPUT (R) 0.246
NPUT 0.260 0.070 0.380 0.450 O.260 0.380
8 OUTPUT (T) 0.600
OUTPUT (R) 0.609
NPUT 0.260 0.50 0.260 0.260 0.320 0.260
9 OUTPUT (T) 0.000
OUTPUT (R) 0.082
INPUT 0.260 0.190 0.450 0.260 0.90 0.190
20 OUTPUT (T) 0.000
OUTPUT (R) 0.085
INPUT 0.380 0.450 0.130 0.580 0.380 0.50
2 OUTPUT (T) .000
OUTPUT (R) 0.936
INPUT 0.90 0.380 0.380 0.380 0.380
OUTPUT (T)
OUTPUT (R)
NPUT 0.070 O.260 0.90 0.260 0.90
23 OUTPUT (T)
OUTPUT (R)
NPUT 0.580 0.60
24 OUTPUT (T)
OUTPUT (R)
38O 0.540 0.320 0.630
INPUT 0.90 0.380 0.320 0.380 0.260

25 OUTPUT (T)
OUTPUT (R)
NPUT 0.380 0.510 0.30 0.32O 0.320
OUTPUT (T)
OUTPUT (R)
BHQXW(TEONVIZLTBHIE)NAG CN
O
L. cN N ?
O CD O O C O
/
FIG 3
INPUT INTERMEDE OUTPUT
SAGE STAGE STAGE
N M OT
OLD DATA
NEW DATA
F. G. 3A
SMOKE SENSOR
SLVs 6
5
4
3
2
F. G. 3B
INPUT INTERMEDE OUTPUT
STAGE STAGE STAGE
N M OT
F. G. 4 400
4
INPUT I/O REL
FOR DEF THROUGH
TEN KEY 8, STORE
IN RAM12
STORE NET WEIGHT
VAL AT CORRES
RAM 13 ADDR
LEARN
AGAIN
N 408
409
G.)
40
SEND DATA SEND-BACK - 41

COM TO n-th DETECTOR
DATA
SENT
BAC 42
Y
READ DATA & 43
STORE IN RAM 1
FIG 5A
O)
INPUT SVS NSLVS6 TO
BLOCK A NPUT STAGE 514
NET STRUCTURE
CALC PROG 700
INPUT SLV-NSLV6 TO 515

BLOCK B NPUT STAGE
FIG 5
NET STRUCTURE 700
CALC PROG
INPUT SLVg NSLVg6 TO 516

414 SET n-th DETECTOR BLOCK C NPUT STAGE
DATA OF RAM N IN
NETCALC
STRUCTURE
PROG |-o
700 NET STRUCTURE
CALC PROG
INPUT O/PUTS OUTs 57
OUTt OUTg OF BLOCKS
A B C TO BLOCK D
NET STRUCTURE
CAL PROG 700
416 DISPLAY FIRE PROB
O/PUTS: OUTS OUTt 518
OUTg
521 E N
Y
F. G. 6 NET STRUCTURE 600

GEN PROG
ALL WEIGHT VAL OF SIG

LINES INTERCON /PUT
INTERMED & O/PUT 601
STAGES ASSUME
CONSTANT
M
2, (Em) IS ASSUME-E0 602
WEIGHT
PRO BETWEEN INTER
MED & O/PUT STAGE
COMPLETE
. WEIGHT N
PRO BE WEEN INTER
ED & O/PUT STAG ALTER WEIGHT VAL
COMPLETE TO Vik -- S
611
N 605
ALTER WEIGHT
617 Y VAL TO will S ALL SET PATTERNS
RETURN 612 ERR GIVEN BY ES
68 ALL SE PATTERNS 606
ERR GIVEN BY ES N <>
63 607
N Y
614 WEIGHT VAL SET
Y TO Vik-S. A
WEIGHT VAL SET
TO Wij-S a DETER OVERAL
ERR ES
615 NDETER oVERALL
ERR ES 609
616 Eo F ES
FG 7
ET STRUCTURE
CALC PROG
701
702
703
704
705
706
707
708
709
710
Y 711
RETURN
FG. 8
26
WHEN m=
(Em) = 0.07655
INPUT STAGE - INTERMEDIATE STAGE
(U33
- 60400
(U34 (U64
2.5785 O - 0357001
- 0- Moamera im Pub womtow an as a to 4.pno us as uses on mana is is a Pop wn wipe we wo
INTERMEDIATE STAGE - OUTPUT STAGE

V1 V21 V31 V41
-6.098 - 1.7389 0.3929 -9.286
HOSNE
?-IV E-t ><*
NJ09 |FŒU,TW/L"
5,168,262
1. 2
due to the influence of noise by the method relying on
FIRE ALARM SYSTEM a table or function.
Accordingly, a second object of the present invention
TECHNICAL FIELD is to provide a fire alarm system having a signal process
The present invention relates to a fire alarm system in ing structure suited for achieving the first object men
which a plurality of physical quantities such as heat, tioned above.
smoke or gases attributable to a fire phenomenon are DISCLOSURE OF THE INVENTION
detected time-serially for thereby making a decision or
judgement as to the occurrence of a fire on the basis of In view of the above objects, there is provided ac
the plurality of time-serial physical quantities. 10 cording to a first mode of carrying out the present in
vention a fire alarm system in which detection informa
BACKGROUND TECHNOLOGY tion output from fire phenomenon detecting means is
In connection with a fire decision made on the basis subjected to signal processing for obtaining a value for
of a plurality of sensor levels that vary with time and are at least one type of fire information, the fire alarm sys
detected time-serially as detection information repre 15 ten comprising:
sentative of physical quantities involved in a fire phe detection information collecting means for collecting
nomenon, there can be conceived a so-called discrimi time-serially a plurality of detection information values
native pattern identification method according to which from the fire phenomenon detecting means, and
a table containing patterns based on a plurality of time signal processing means for performing signal pro
serial sensor levels together with fire information for
20 cessing on the basis of the plurality of detection infor
each of the patterns is prepared and stored in a ROM or mation values collected time-serially from said fire phe
the like, wherein the pattern information in the table is nomenon detecting means by said detecting information
compared with time-serial sensor levels actually de collecting means by correspondingly imparting weights
tected, for thereby allowing the fire decision to be to each input time-serial detection information value, in
made.
25 accordance with degrees of contribution thereof to said
Further, it is also conceivable to define a function fire information upon input of said time-serial detection
having as variables the values of a plurality of time information values, for thereby allowing the fire infor
serial sensor levels, wherein the fire decision is made on mation value to be arithmetically determined on the
the basis of input/output relations with the aid of the basis of the weighted detection information values.
function.
30 Further, according to a second mode for carrying out
In any case, the decision as to whether or not a fire is
the present invention, there is provided a fire alarm
occuring is based on the detected sensor levels. In this system in which detection information output from a
plurality of fire phenomenon detecting means is sub
conjunction, it is extremely desirable if the fire monitor jected to signal processing for obtaining a value for at
ing operation can be effectuated with highly improved 35 least one type of fire information, the fire alarm system
accuracy by virtue of the capability to finely and thor comprising:
oughly monitor fire phenomena inclusive of smoldering detection information collecting means for collecting
fires and flaming fires while making available the infor the detection information from each of the fire phenom
mation concerning the possibility of a fire, i.e. fire prob enon detecting means while collecting a plurality of
ability and the level of danger as well as the capability time-serial detection information values from at least
of eliminating the possibility of false alarm generation one of said fire phenomenon detection means, and
due to noise or other causes. signal processing means for performing the signal
Accordingly, a first object of the present invention is processing on the basis of said plurality of detection
to provide a fire alarm system for making a decision as information values collected from said plurality of fire
to the occurrence of a fire on the basis of a plurality of 45 phenomenon detecting means through said detection
sensor levels detected time-serially, a system which is information collecting means by correspondingly im
not only capable of making a decision as to the occur parting weights to each detection information value as
rence of a fire but also capable of finely and thoroughly input, in accordance with degrees of contribution
monitoring the fire probability and the level of danger thereof to said fire information upon inputting of said
as well as fire phenomena inclusive of smoldering fires 50 detection information values, for thereby allowing said
and flaming fires with regard to such situations or states fire information value to be arithmetically determined
which may lead to a fire while eliminating the possibil on the basis of the weighted detection information val
ity of erroneous or false alarm generation from the eS.
influence of noise or the like. In conjunction with a second mode for carrying out
In case fire information corresponding to a plurality 55 the invention, the signal processing means may be so
of time-serial sensor levels should be defined in a table implemented that the detection information values col
stored in a ROM or the like as described above in an lected by the detection information collecting means
effort to accomplish the above object, an increase in the can be input en bloc to the signal processing means
number of input points or data would involve a more whereon the latter correspondingly weights the input
explosive increase in the number of combinations of 60 detection information values for arithmetically deter
such inputs, requiring prodigious labor and a large ca mining the fire information value, or the signal process
pacity ROM table for describing all the combinations, ing means may include first auxiliary processing means
which would be practically impossible. Further, de providing in correspondence with said at least one fire
scription of the input/output relations in terms of the phenomenon detecting means by which said plural
functions as mentioned above is also practically impos 65 time-serial detection information values are collected,
sible because of the limitation encountered in expressing for performing an arithmetic operation to obtain indi
such complex relations, not to say of the elimination of vidual fire information values, and second auxiliary
the possibility of erroneous or false alarm generation processing means for processing the individual fire in
3
5,168,262
4.
formation values input from said first auxiliary process area are adjusted by said adjusting means on the basis of
ing means and detection information values input from the contents of said table.
the fire phenomenon detecting means which but col Although this kind of storage means can be previ
lects, not time-serially a detection information value, to ously prepared at the manufacturing stage or at other
thereby derive the final fire information having highly appropriate times for subsequent use, it may initially be
enhanced reliability. created internally of the fire alarm system upon initial
According to a third mode for carrying out the pres ization thereof. In case the storage means is created
ent invention, there is provided a fire alarm system for internally of the fire alarm system, the table and the
obtaining a value for at least one type of fire information adjusting means are also incorporated in the fire alarm
by processing signals representative of detection infor 10 system.
mation outputs from a plurality of fire phenomenon The adjusting means adjusts the weight values to be
detecting means, which system comprises: stored in the storage means such that the difference
detection information collecting means for collecting between a fire information value output from the signal
time-serially a plurality of detection information values processing net and the input/output value listed in the
from each of said fire phenomenon detecting means; 15 definition table is minimized. Once the storage means
and has been prepared in this manner, the signal processing
signal processing means for performing signal pro means or the auxiliary signal processing means can
cessing on the basis of the detection information values perform an arithmetic operation by using the weight
collected by said detection information values collect values stored in the storage means to thereby output the
ing means from said plurality of fire phenomenon de 20 desired output values for all the input values. Thus, the
tecting means by imparting corresponding weights to signal processing means or the auxiliary signal process
each of said detection information values upon input ing means can cope with combinations of a plurality of
ting thereof in accordance with an extent to which said time-serially detected information values which are not
each input detection information value contributes to defined in the definition table, whereby the values rep
said fire information and arithmetically determining 25 resentative of the desired fire information (fire probabil
said fire information value on the basis of the weighted ity, the level of danger, probability of the smoldering
detection information values. fire, etc.) can be indicated. In this manner, a finer fire
In conjunction with a third mode for carrying out the decision can be made on the basis of the time-serially
invention, the signal processing means may be so imple detected information values collected by the detected
mented that the detection information values collected
30 information collecting means.
by the detection information collecting means can be As can be appreciated from the above, by using the
input en bloc to the signal processing means, whereon storage area storing the weight values and the signal
the latter correspondingly weights the input detection processing means (or auxiliary processing means), it is
information values for arithmetically determining the unnecessary to define all the pattern combinations but is
fire information value, or the signal processing means
35 sufficient to define the combinations only for the impor
may include first auxiliary processing means provided tant points or locations when defining the input/output
in correspondence with said fire phenomenon detecting relations. Further, when the necessity arises for describ
ing in detail among others those regions including a
means by which said plural time-serial detection infor singular point or maximum or minimum point where the
mation values are collected, for performing arithmetic output values change remarkably even for a small devi
operation to obtain individual fire information values, ation in the input value, then such regions and peripher
and Second auxiliary processing means for processing ies thereof may be defined finely with other regions
the individual fire information values input from said being defined roughly.
first auxiliary processing means to thereby derive final When an input/output relation is to be changed, this
fire information having enhanced reliability. 45 can be achieved either by defining an output value for
In any one of the abovementioned modes for carrying an input value differing from that defined previously or
out the invention, the signal processing means should by creating a new definition for a region not yet de
preferably include storage means for previously storing fined. In this conjunction, it is noted that such alteration
weight values for correspondingly weighting the infor of the definition can be readily realized in the form of
mation values, respectively. The weight values stored in 50 modification of the weight values by running the adjust
the storage means are so selected or established as to ing means (net structure generating program). In other
cause the fire information value arithemtically deter words, by altering the definitions, it is possible to accu
mined by said signal processing means in response to the rately realize a decision or judgement concerning a fire,
input of a particular set of the information values to danger, etc.
approximate desired fire information value which is to 55 In any one of the modes for carrying out the inven
be derived from said particular set of the information tion, the practical embodiment of the signal processing
values. means or auxiliary processing means should preferably
In conjunction with the preparation of the storage be so implemented as to perform the arithmetical deter
means, there may be provided a table for storing therein mination hierarchically, in which instead of straightfor
a particular set of information values together with at wardly calculating the fire information value from a
least one fire information value which is to be obtained plurality of detection information values collected by
when said particular set of information values is given the detection information collecting means, interim or
and adjusting means for adjusting the weights so that intermediate information values is once determined
said fire information value arithmetically determined by arithmetically from the information values as input,
said signal processing means when said particular set of 65 whereon the fire information value is arithmetically
information values stored in said table is supplied can determined from the intermediate information values.
approximate said fire information value stored in said Such hierarchical structure may be realized in stages
table, wherein said weight values stored in said storage comprising a plurality of intermediate layers, in each of
5,168,262
5 6
which layers a desired number of intermediate informa sensor levels on the assumption that the weight values
tion values to be arithmetically determined may be are established as shown in FIG. 8.
established. By way of example, in the case of a two BEST MODE FOR CARRYING OUT THE
stage hierachical structure including an input-inter INVENTION
mediate section and an intermediate-output section, the 5
intermediate information values are once determined In the following, the present invention will be de
arithmetically from the input detection information scribed in conjunction with exemplary embodiments
values, whereon the fire information value to be output thereof.
is determined arithmetically on the basis of the interme FIG. 1 is a block circuit diagram showing a so-called
diate information values. In that case, initial weights are 10 analogue type fire alarm system to which an embodi
imparted separately for each of the input information ment of the present invention is applied and in which
values before deriving the intermediate information sensor levels representative of analogue physical quanti
values, which is then followed by second weighting of ties inherent or attributable to the fire phenomena as
the individual intermediate information values, respec detected by individual fire detectors are sent out to
tively. In this manner, the fire information value can be 15 receiving means such as a receiver, repeater or the like,
determined as the output information. The values of the wherein the receiving means is adapted to make a deci
individual intermediate information plays no important sion as to the occurrence of a fire on the basis of the
role. The signal processing means may initially be ad sensor levels as collected. However, it goes without
justed upon initialization processing thereof or at any saying that the present invention can equally be applied
approate time point in a manufacturing process in re- 20 to an on/off type fire alarm system in which the deci
spect to the first and second weight values by the afore sion as to the occurrence of the fire is made at the indi
mentioned adjusting means. vidual fire detectors, wherein only the results of the
When the fire alarm system comprises a receiving decision are sent to the receiving means.
part such as a fire control panel and a plurality of fire In FIG. 1, reference character RE denotes a fire
detectors connected to the receiving part and each 25 control panel, and DE 1 to DEN designate N analogue
including at least one fire phenomenon detecting means type fire detectors connected to the fire control panel
for detecting a physical quantity attributable to the fire RE by way of a transmission line L which may be con
phenomenon, the abovementioned signal processing stituted, for example, by a pair of conductors serving for
means may be incorporated either in the receiving part both electric power supply and signal transmission, in
or in the fire detectors. When the signal processing 30 which only one of the fire detectors is illustrated in
means includes auxiliary processing means, a certain detail with respect to the internal circuit configuration.
one or ones of the auxiliary processing means may be In the fire control panel RE:
provided in the fire detectors while the remaining auxil MPU1 denotes a microprocessor:
iary processing means may be provided in the receiving ROM11 denotes a program storage area for storing
part. 35 programs relevant to the operation of the inventive
BRIEF DESCRIPTION OF THE DRAWINGS system which will be described hereinafter;
ROM12 denotes a constant table storage area for
FIG. 1 and FIG. 1A are block circuit diagrams show storing various constant tables containing criteria and
ing fire alarm systems according to first and second other information for discriminative identification of
exemplary embodiments of the present invention, re- 40 the fires for all of the fire detectors;
spectively; ROM13 denotes a terminal address table storage area
FIGS. 2 and 2' and FIG. 2A are views showing defi for storing addresses of the individual fire detectors;
nition tables employed in the first and second embodi RAM11 denotes a work area;
ments of the present invention, respectively, each con RAM12 denotes a definition table storage area for
taining defined inputs "INPUT” and defined outputs 45 storing definition tables for all of the fire detectors, as
"OUTPUTCT)' together with actually measured fire will be described hereinafter;
information values "OUTPUT(R)' output from the net RAM13 denotes a weight value storage area for stor
structure in response to a supply of the defined inputs ing weight values of signal lines for all of the fire detec
INPUT: tors, as will be described later;
FIG.3 and FIGS. 3A and 3B are views for conceptu- 50 TRX1 denotes a signal transmission/reception part
ally illustrating signal processing nets employed in the which is constituted by a serial-to-parallel converter, a
first and second embodiments of the present invention, parallel-to-serial converter, etc.;
respectively; DP denotes a display such as a CRT or the like;
FIG. 4 is a flow chart for illustrating operations of the KY denotes a ten key for inputting data for teaching,
systems shown in FIG. 1 and FIG. 1A; 55 as will be described hereinafter; and
FIG. 5 and FIG. 5A are flow charts for illustrating IF11, IF12 and IF13 denote interfaces, respectively.
operations of the systems shown in FIG. 1 and FIG. 1A, Further, in connection with the fire detector DE:
respectively; MP2 denotes a microprocessor;
FIG. 6 is a flow chart for illustrating a net structure ROM21 denotes a program storage area;
generating program (weight value adjusting means) 60 ROM22 denotes an own address storage area;
shown in FIG. 4; RAM2 denotes a work area;
FIG. 7 is a flow chart for illustrating the net structure FS denotes a fire phemonenon detecting means for
calculation programs shown in FIG. 5 and FIG. 5A; detecting physical quantities such as those of heat,
FIG. 8 is a view showing individual weight values smoke, gases or the like ascribable to a fire phenomena,
used in obtaining the actually measured values of the 65 which means is composed of a smoke sensor of the
fire information shown in FIG. 2; and scattered light type in the case of the instant embodi
FIG.9 is a view showing fire probability output from ment. The smoke sensor part FS includes a light emit
the net structure in response to actual changes in the ting circuit, a light receiving circuit, a dark box of laby
5,168,262
7 8
rinth structure, an amplifier, a sampling and hold cir Sn (1sms5). The values of Sn can be expressed in
cuit, an analogue-to-digital converter and others, al accordance with the values of SLV and SLV-1, as
though they are not shown; follows:
TRX2 denotes a signal transmission/reception part When SLV,203 and when SLV,s SLV, 1,
similar to TRX1; and
IF21 and IF22 denote interfaces. with "Siv.203 and when SLV-1>SLVn,
In precedence to the concrete description of the oper
ation of the exemplary embodiment of the present in we'lv.203 and when SLVSSLV-1,
vention which will be made later, a description will first
be directed to the concept underlying the illustrated O
wilv.<0. and when SLVDSLV-1,
embodiments. ru.
With the instant exemplary embodiment, it is contem Accordingly, the fire probability over all of the five
plated to allow various fire decisions such as the proba intervals is given by
bility of a fire and the degree or level of danger to be
made rapidly and correctly on the basis of a plurality of 15 5
sensor levels supplied time-serially from the sensor parts S = insX (Sn).
which detect the physical quantities of the fire phenom
enon. To this end, the sensor levels from the sensor part The overall fire probability S determined in the man
sampled every fifth second are collected over a period ner described above provided the base for deriving the
of twenty-five seconds, wherein the six sensor levels in 20 values enumerated at the intermediate rows labeled
total are input to a net structure as a pattern to thereby “OUTPUTCT) in the definition table shown in FIG. 2.
allow the probability of a fire to be obtained as the However, all the values thus determined are not utilized
output of the net structure, the operation of which will intact as the values of "OUTPUT(T)', but instead the
first be described by reference to FIGS. 2 and 3. values most approximating the actual values are en
FIG. 2 shows a definition table which defines a true 25
or highly accurate probability for 26 types of combina ployed with the influence of noise, statistical data reli
tions or patterns of the six sensor levels, in which for ability and other factors in the environment where the
each of the patterns numbered up to the 26-th pattern, sensors are installed taken into consideration. Further,
six time-serial sensor levels are shown at the uppermost for sensor levels not varying linearly, as can be seen in
row labeled "INPUT'. Of these six sensor levels, the 30 the patterns Nos. 20 to 26, similar definitions are
leftmost one corresponds to the level sampled twenty adopted to ensure redundancy so as to sufficiently and
five seconds before, wherein the data subsequently sam elastically cope with the actual time-serial sensor level
pled sequentially are shown serially in the direction patterns. For example, in the case of the pattern No. 5,
from left to right as viewed in the figure. Accordingly, the output "OUTPUT(T)" assumes a value of "0.800”
the rightmost data represents the sensor level sampled 35 which should be "0.7' in accordance with the concept
last. At the intermediate row labeled "OUTPUT(T)' described above. This can be explained by the fact that
for each of the patterns numbered, there are enumerated the sensor level SLV5="0.380” is ascribable to the
the probability of a fire in terms of numerical values in influence of noise because only the sensor level SLV5
a range of "0" to "1" in association with the six sensor falls extremely while the levels preceding and succeed
levels at the upper row, respectively. The sensor levels ing to the sensor level SLVs increases. Accordingly, in
at the upper row are also given in terms of numerical reality, the sensor level SLV5 is considered to lie within
values obtained through conversion or transformation a range of SLV4<SLV5<SLV6. By taking this into
processing. By way of example, the sensor levels of "O" account, “O.800” is placed at "OUTPUTCT)".
to "1" correspond to the smoke concentrations in a This kind of definition table can be prepared precisely
range of 0 to 20%/n detected by a smoke sensor. At the 45 on the basis of the concept described above and through
lower row labeled "OUTPUT(R)', there are shown the experiments performed at places where the fire detec
values of the probability of the fire measured actually, tors are installed while taking into consideration the
as will be described later. characteristics of the fire detectors and the statistical
The probability "OUTPUTCT)' to be obtained when reliability of data. It is, however, practically impossible
a single pattern of the six sensor levels shown in FIG. 2 50 to prepare this sort of table for all the patterns let alone
is given can be derived generally on the basis of the the twenty-six combinations of the six sensor levels. In
concept which will be described below. contrast, according to the teachings of the present in
When the sensor level converted to a nunnerical value vention described subsequently, it is possible to deter
in the range of “0” to “1” exceeds "0.3' and when it is mine accurately the fire probability for all the patterns
maintained constant or tends to increase, "0.2' is added 55 on the basis of the six time-serial sensor levels with the
to the value of the fire probability per interval. On the filtering effect against noise, etc. being taking into ac
other hand, when the sensor level which exceeds "0.3' Count.
tends to decrease currently, "0.1' is added to the fire Now, for convenience of elucidation of the teachings
probability per interval. At all other intervals, "0" is of the present invention, a net structure such as illus
added to the fire probability. The sum of these fire trated in FIG. 3 will be utilized. The object of this net
probabilities based on the six sensor levels and added structure, is to obtain the precise fire probability by
together over all of the five intervals is utilized as the supplying six sensor levels to the net structure on the
overall fire probability. assumption that such net structures are incorporated in
The above will be explained with the aid of expres the fire probability RE in correspondence with the
sions on the assumption that a certain sensor level is 65 individual fire detectors DE1 to DEn, respectively. In
represented by SLV, the sensor level sampled five the net structure shown in FIG. 3, N1 to IN6 indicated
seconds later is represented by SLV-1, and that the on the left-hand side will be referred to as the input
fire probability ratio at each interval is represented by stage layers, while OT indicated on the right-hand side
5,168,262
10
is referred to as the output layer or stage OT. There are
input to the six input layers IN1 to IN6 the six sensor I .. 1
levels each converted to numerical values in the range NET,0) = 2 (IN. Wi) (Eq. 1)
of "0" to “1”. On the other hand, there is output from 5
the output layer OT the fire probability represented by When the value NET (j) is converted to a value in a
a numerical value from "O' to "1'. Further, four layers range of "0" to "1" with the aid of the sigmoid function,
IM1-IM4 shown, only by way of example, are referred for example, which is then represented by IMj, the
to as intermediate stage layers, respectively. These in following relation applies valid:
termediate stage layers IM1-IM4 receive the signals 10
from the individual input stage layers IN 1-IN4 and out IMG = 1 + EXP-NET(j) y) (Eq. 2)
put the signals to the output stage OT1. It is assumed
that the signals travel from the input stage to the output
stage without traveling in the opposite direction and Similarly, the total sum NET2 (k) of the inputs to the
without undergoing signal-coupling among the layers 15 output stage OTk can be expressed by:
of the same stage. It is additionally assumed that no J .. 3
direct signal coupling is made from the input stage lay NET(t) = (IM). Vik) (Eq. 3)
ers to the output stage. Accordingly, there exist twenty
four signal lines extending from the input stage to the
intermediate stage. Similarly, four signal lines extend 20 When the value NET2 (k) is converted to a value in the
from the intermediate stage to the output stage. range of "0" to “1” by the sigmoid function, which is
The signal lines shown in FIG. 3 have respective then represented by OTk, the following relation applies
weight values or coupling degrees which vary in depen valid:
dence on the values to be output from the output stage 25
in response to the signals input at the input stage, OTk = (Eq. 4)
wherein signal transmission capability of the signal line 1 EXP-NET.(6).
is increased as the weight value thereof increases. The In this manner, the relations between the input values
weight values of the twenty-four signal lines between IN1 - IN6 and the output value OT can be represented
the input stage and the intermediate stage, as well as the 30 by the expressions Eq. 1 to Eq. 4 by using the weight
four signal lines between the intermediate stage and the values. In the above expressions, y1 and y2 represent
output stage, and thus the weight values of twenty-eight adjustment coefficients of the sigmoid curve. In the case
signal lines in total, are stored in the weight value stor of the instant embodiment, they can be appropriately
age area RAM13 shown in FIG. 1 at the areas allocated selected such that y1 = 1.0 and y = 1.2. By using these
to the individual fire detectors, respectively, after hav 35 adjustment coefficients, it is possible to adjust the incli
ing been initially adjusted in accordance with the rela nation of the sigmoid curve to thereby regulate the
tions between the inputs and the outputs. The weight convergence rate for reducing errors.
values thus stored are subsequently, made use of in the In preparing the net structure generating program,
fire monitoring operation. A. one of the twenty-six patterns or combinations of the six
In more concrete terms, the six values at the upper sensor levels shown in the definition table stored in the
row “INPUT” for each of the pattern numbers (Nos.) in storage area RAM12 is input to the input stage layers
the definition table shown in FIG. 2 are supplied to the IN-IN6, whereon the value of OTk(where k = 1 in the
input stage layers IN1 to IN6, respectively, in accor case of the instant embodiment) output from the output
dance with a net structure generating program which 45
stage as the result of the calculations according to the
will be described hereinafter, wherein the value output expressions Eq. 1 to Eq. 4 mentioned above is compared
from the output layer OT1 in response to the inputs with the teacher signal outputs T shown at the interme
mentioned above are compared with the fire probabili diate row in FIG. 2. At that time, any error Em which
ties T1 listed at the intermediate row "OUTPUT(T)” in may occur at the output stage (where m = 1 - M and
the table shown in FIG. 2 and serving as the teacher 50
M=26 in the case of the instant embodiment) is repre
sented by the following expression:
signals or the data for learning, and the weight values of
the individual signal lines are altered so that the error or
difference resulting from the comparison are reduced to En = , (OTk - T) (Eq. 5)
a minimum. In this manner, data very closely approxi
mating all of the functions shown in the definition table 55
of FIG. 2 for only twenty-six combinations or patterns where OT1 represents the value determined in accor
dance with the expression Eq. 4 mentioned hereinbe
can be taught in the net structure shown in FIG. 3. fore. The value E totaling the error Em for all the M
Now assuming that the weight value between the patterns or combinations, i.e. the twenty-four combina
input stage layer INi and the intermediate stage layer tions contained in the table of FIG. 2 is given by:
IMj is represented by Wij with the weight value be
tween the intermediate stage layer IMj and the output M (Eq. 6)
stage OTk being represented by Vijk (where i=1 ~I, E = nx= t1 (Em)
j=1-J and where K = 1 with I=6, J = 4 and K= 1 in
the case of the instant embodiment) and further assum 65 Finally, an operation is performed for adjusting the
ing that each of the weight values Wij and Vjk can take weight values of the signal lines on a one-by-one basis so
positive, zero or negative values, the total sum NET (j) that the value E given by the expression Eq. 6 is mini
of the inputs to the intermediate stage IMj is given by mized. The weight values stored in the fire detector
5,168,262 12
11
area of the storage area RAM13 are updated with these Referring to FIG. 4, the net structure generating
new weight values to be utilized in the ordinary fire program is executed sequentially for each of the N fire
monitoring operation. The adjustment of the weight detectors, starting from the No. 1 fire detector.
values for the signal lines as described above is per Describing the operation of the net structure generat
formed for all the fire detectors included in the fire ing program for the n-th fire detector (n = 1 - N), the six
alarm system. sensor levels listed at the upper row and the fire proba
Upon completion of the teaching of the table contents bility at the intermediate row in the definition table
shown in FIG. 2 for the net structure illustrated only described previously by reference to FIG. 2 are first
conceptually in FIG. 3, i.e. upon completion of the given as the teaching inputs or the inputs for learning
adjustment of the weight values of the signal lines on a 10 through the learning data input ten key KY (step 404).
line-by-line basis, the actual fire monitoring operation is Although a definition table is prepared for each of the
then performed by determining through calculation fire detectors in view of the fact that the environments
with the aid of a net structure calculation program where the fire detectors are installed and the character
(which will be described hereinafter) the value obtained istics thereof differ from one to another fire detector, it
from the output stage OT in response to the input of goes without saying that a similar definition table can be
the six sensor levels sampled time-serially over the per used for those fire detectors having similar characteris
iod of twenty-five seconds to the input stage of the net tics and similar environmental conditions.
structure in accordance with the expressions Eq. 1 to When the contents of the definition table for the n-th
fire detector are stored in the n-th fire detector area
Eq. 4 mentioned above, wherein the fire decision is provided in the definition table storage area RAM12
20
made by comparing the values resulting from the above through the ten key (when Y results from step 403),
calculation with the reference value of the fire probabil then processing proceeds to the execution of the net
ity.
In the foregoing description, it has been assumed that structure
FIG. 6.
generating program 600 also illustrated in
the number of information values input to the input 25 In the first place, the weight values Wij and Vik of
stage layers is six with that of the information values the twenty-eight signal lines in total, including 24 lines
output from the output stage being one. It goes, how provided between the input stage and the intermediate
ever, without saying that the number of input informa stage and 4 lines provided between the intermediate
tion values as well as of the output information values
can be selected arbitrarily, as occasion requires. As the 30 stage and the output stage as described hereinbefore in
conjunction with FIG. 3, are set at given constant val
information values output from the output stage, there ues, respectively, (step 601). Subsequently, on the basis
can be mentioned in addition to the fire probability of the weight values set to be constant, the totaled value
other various information values such as the degree or (E of the expression Eq. 6) of the squares of errors
level of danger, the concentration of smoke, seethrough between the output values OT and the teacher output
or visible distance, etc. 35 values T are determined in accordance with the previ
Further, although it has also been assumed that there ously mentioned expressions Eq. 1 to Eq. 6 for all the M
is one intermediate stage that includes four elements, combinations (M=26 in the case of the illustrated em
the relation between the number of the elements in bodiment) listed in the definition table of FIG. 2,
cluded in one intermediate stage and those of the input wherein the result as obtained is represented by Eo (step
information values and output information values is 602).
generally such that when the number of input informa Next, an operation is performed to adjust one by one
tion values is increased, the number of elements in the weight values of the four signal lines between the
cluded in the intermediate stage should preferably be intermediate stage and the output stage so that the over
increased correspondingly in order to minimize error. all error value E is minimized for inputting the same
Of course, by increasing the number of intermediate 45 definition table (N of step 603). Because the adjustment
stages, the accuracy is further improved. of the weight values is made only for the signal lines
Further, it has been described that the total sun extending between the intermediate stage and the out
NET (j) of the inputs to the individual elements at the put stage, no changes can take place in the values deter
intermediate stage as calculated in accordance with the mined in accordance with the expressions Eq. 1 and Eq.
expression (Eq. 1) is converted to a value in the range of 50 2. At first, the weight value V 1 of the first one signal
"0" to “1” with the aid of the sigmoid function, wherein line is altered to a weight value of V1 1--S (step 604)
the value thus obtained is used in the expression (Eq. 3). and the calculations are performed similarly in accor
It should, however, be mentioned that in place of the dance with the expressions Eq 3 to Eq. 6. The final
conversion of NET1 () to a value of "0" to "1", NET error value E determined from the expression Eq. 6 is
(i) can be directly used in the expression (Eq. 3) in place 55 represented by Es (step 605). Then, the value of Es is
of IM. Even in that case, the final output information compared with the overall error value E before alter
value is converted to a value in the range of "0" to "1" ing the weight value (step 606).
(Eq. 4) to be output from the output stage OT1. If EssBo (N of step 606), the value E is set as a new
In the illustrated embodiment, neither the elements or value of Eo (step 609), while the updated weight value
layers at the intermediate layer stage are mutually cou of (V 1--S) is stored at an appropriate location in the
pled, or are the elements of the input and output stages work area.
mutually coupled. Nevertheless, the object of the pres On the other hand, when Ed Eo (Y of the step 606),
ent application can be accomplished by altering the this means that the direction in which the weight value
weight values in such a sense that error is reduced. has been changed is erroneous. Accordingly, the weight
FIG. 4 to FIG. 7 are flow charts for illustrating oper 65 value is altered in the opposite direction starting from
ations of the inventive system executed in accordance the original weight value V11, being then followed by
with programs stored in the storage area ROM1 shown the calculation of Es by using a weight value of V1
in F.G. 1. 1-S-É in accordance with the expressions Eq 3 to Eq.
13
5,168,262
14
6 (steps 607, 608), wherein the value of E thus deter set initially in terms of the teacher signals. The corre
mined is set as the new value of Eo (step 609), while the sponding weight values for the actually measured val
altered weight value of V1 1-S-É is stored at an appro ues "OUTPUT(R)' of the fire probability are shown in
priate location in the work area. FIG. 8.
It should be mentioned that (3 represents a coefficient 5 FIG. 9 illustrates graphically the actually measured
proportional to Es-Eo and that S is variable as a values of the fire probability output from the net struc
function of the number of times the weight value is ture upon the input thereto of the real arbitrary values
altered or changed and assumes a smaller value as said of the sensor levels varying from time to time in addi
number of times increases. tion to the specific patterns of the six sensor levels,
After completion of the alteration and adjustment of O wherein time is taken along the abscissa while there is
V11 through the steps 604 to 609, then the alteration and taken along the ordinate the sensor level SLV varying
adjustment of the weight values V2 1 to V4 for the from time and the fire probability F output from the net
remaining three signal lines are sequentially performed structure.
through the similar processing steps 604 to 609. By defining the time-serial input information values
Upon completion of the adjustment of the weight 15 of the six sensors and the fire probability serving as the
values Vijk for all the signal lines extending between the teacher signal in terms of twenty-six patterns in the
intermediate layer stage and the output stage in this way manner mentioned above, those combinations of the
(Y of the step 603), a similar adjustment is next per sensor outputs which are not contained in the definition
formed on the weight values Wij for the signal lines table can also be determined through interpolation by
between the input stage and the intermediate stage at 20 the net structure, whereby the optimum output is pro
steps 610 to 616 all in accordance with the expressions duced as the indication or answer. In the case of the
Eq. 1 to Eq. 6 so that any error can be minimized.
When the adjustment of the weight values for all the instant embodiment, it is assumed that the numbers of
signal lines has been completed (Y of step 610), the the inputs and the outputs to and from the net structure
value Eo having been reduced in this way is compared 25 are six and one, respectively. However, it can readily be
with a predetermined value C. When the former is still understood by those skilled in the art that the sensor
greater than the value C (N of a step 617), the step 603 input number as well as the sensor output number can
is regained for diminishing further the error, wherein be increased or decreased, as occasion requires. Besides,
the procedure for adjustment of the weight values be there may be conceived as the output information a
tween the intermediate stage and the output stage 30 variety of combinations inclusive of the probability of
through the steps 604 to 609 described above is repeated there being no fire, visibility or see-through distance,
again. When the value E becomes equal to or smaller walking speed, probability of fire extinguishing and
than the predetermined value C after the repeated ad others.
justment (Y of step 617), the processing proceeds to a When the adjustment of the weight values for the
step 406 shown in FIG. 4, where the altered and ad 35 signal lines has been performed for all of the N fire
justed individual weight values Vik and Wij for the detectors incorporated in the fire alarm system (Y of a
twenty-eight signal lines are stored in the associated step 407) and when it is decided that there is no neces
n-th fire detector area of the storage area RAM13 at the sity for the repeated learning (N of step 408), then the
corresponding addresses, respectively. fire monitoring operation of the fire detectors is acti
Through the operation described above, the values of 40 vated sequentially, starting from the first fire detector.
S, a, g, C, etc. are stored in the storage area ROM12 for Describing the fire monitoring operation in connec
the various constants table. tion with the n-th fire detector DEn, a data send-back
Since the final error value of Eo can not assume zero, command for the n-th fire detector DEn is sent out onto
the adjustment of the weight values for the signal lines the signal line L from the signal transmission/reception
has to be terminal at an appropriate value. In this con 45 part TRX1 through the interface IF11 (step 411).
junction, it is noted that in addition to the termination of Upon reception of the send-back command by the
the adjustment at the time point when Eo becomes equal n-th fire detector DEn, the latter reads through the
to or smaller than C, as indicated at the step 617, it is interface IF21 the sensor level (based on such physical
also possible to previously determine the number of quantities as smoke, heat or gases) detected by the sen
times the adjustment of the weight value is to be per 50 sor part, i.e. the fire phenomenon detecting means FS
formed, wherein the adjustment is automatically ended and converted into digital quantities by means of the
when the said predetermined number of times has been incorporated analogue-to-digital converter with the aid
attained. of a program stored in the program storage area
The values at the lower row "OUTPUTOR)' in each ROM21 and sends out the sensor level from the signal
of the patterns numbered indicate the fire probability 55 transmission/reception part TRX2 through the inter
output from the net structure as OT in response to the face IF22.
six sensor levels SLV1 - SLV6 indicated at the upper Upon reception of the send-back data from the sensor
row in FIG. 2 and supplied to the net structure as IN, part of the n-th fire detector DEn (Y of a step 412), the
wherein the net structure is so realized as to repeat the sensor levels as sent back are stored in the work area
adjustment at the steps 603-616 until the expression RAM11 (step 413).
(Eq. 6) has assumed the following value: In the work area RAM11, areas are allocated for
storing a plurality of sensor levels for the individual fire
E= (Em) = 0.07655 detectors, respectively, so that the sensor levels sent
n back from the fire detectors upon every polling are held
65 for a predetermined time with the oldest data or sensor
It will be seen from FIG. 2 that the fire probability level being discarded. For example, assuming that the
"OUTPUT(R)' actually output from the net structure period taken for polling each of the fire detectors
approximates very closely the values of "OUTPUT(T)" DE1-DEN is five seconds with the abovementioned
5,168,262
15 16
predetermined period thus being twenty-five seconds, the factory, wherein the weight values would be stored
then the sensor levels obtained through six times of in the EPROM at step 406, the EPROM then being
polling are constantly stored for each of the fire detec mounted on the fire detector. For the fire detector, the
tOS. processing including step 409 shown in FIG. 4 to step
When the sensor level sent back from the n-th fire 418 in FIG. 5 is executed.
detector DEn is stored at the area assigned to the n-th In the following, a description will be made of an
fire detector of the work area RAM11 with the oldest other preferred embodiment of the present invention by
data being discarded (step 413), then the six sensor lev referring to FIGS. 1A, 2, 2A, 3, 3A, 3B, 4, 5A, 6 and 7.
els stored in the area assigned to the n-th fire detector At first, it should be mentioned that those drawings
are converted, respectively, to the numerical values INi 10 showing the second exemplary embodiment that are the
(where is 1-6) in the range of “0” to “1” to be input to same type as those referred to in the description of the
the net structure calculation program (step 414), first embodiment are labeled with the same figure num
wherein the net structure calculation program 700 bers as those used in conjunction with the first embodi
shown in FIG. 7 is executed. ment but with a suffix of A or B. Further, since FIGS.
Through the net structure calculation program 700, 5 2, 3, 4, 6 and 7 remain the same as in the case of the first
NET1 () is arithmetically determined in accordance embodiment, these figures are referred to as they are,
with the expression Eq. 1 mentioned hereinbefore (step without being suffixed with an A or B.
703), the resulting value then being converted into the FIG. 1A shows in a block circuit diagram a so-called
value IMjin accordance with the expression Eq. 2 (step analogue type fire alarm system to which the present
704). When the IMj value is determind for all IM1 to 20 invention is applied and in which sensor levels repre
IMJ (where J =4) (Y of step 705), then NET2 (k) is senting the physical quantities produced by the fire
calculated by using the value of IM in accordance with phenomena and detected by the individual fire detec
the previously mentioned expression Eq. 3 (step 708), tors are sent to a receiving means such as a control
the values resulting from the calculation then being panel, repeater or the like, wherein the receiving means
converted into the values of OTk (where k=1 --K) 25 is adapted to make the decision concerning the occur
(step 709). When the value of OTk, i.e. the value of the rence of a fire on the basis of the sensor levels as co
fire probability OT1 has been determined (Y of a step lected. Of course, it goes without saying that the inven
710), the processing illustrated in the flow chart of FIG. tion can equally be applied to an on/off type fire alarm
5 is regained. Now, referring to FIG. 5, the value of system in which the fire decision is performed at the
OT1 is displayed, as it is, as the fire probability (step 415) 30 individual fire detectors with only the results of the
and compared with the reference value A of the fire decision being sent to the receiving means.
probability read out from the various constant table In FIG. 1A, reference character RE' denotes a fire
storage area ROM12 (step 416). When OT12A, the fire control panel, and DE1' to DEN designate N analogue
indication is activated (step 417). type multi-element fire detectors connected to the fire
Through the procedure described above, the fire 35 control panel RE' by way of a transmission line L
monitoring operation for the n-th fire detector comes to which may be constituted, for example, by a pair of
an end, wherein a similar fire monitoring operation is conductors serving for the electric power supply and
performed for the next fire detector. the signal transmission, in which only one of the fire
Although it has been described in conjunction with detectors is illustrated in detail in respect to the internal
the above embodiment that the data is artificially input circuit configuration. Parenthetically, it should be men
to the definition table storage area RAM12 to thereby tioned that not all of N fire detectors are necessarily
allow the weight values to be stored in the storage area multi-element fire detectors and a plurality of different
RAM13 on the basis of the input data through the net types of fire detectors may be combined to form one
structure generating program, it is equally possible to multi-element fire detector. Accordingly, with the ex
determine the weight values by using the net structure 45 pression "n-th fire detector (n=1-N)' used in the
generating program at a manufacturing stage in a fac following description, it is intended to cover both single
tory and store the weight values in a ROM such as an multi-element fire detectors and a set including a plural
EPROM or the like, the ROM being then incorporated ity of different types of single-element fire detectors.
in the system. The fire control panel RE' has a structure corre
In place of the analogue type fire alarm system de 50 sponding to that of the fire control panel RE shown in
scribed above in conjunction with the exemplary em FIG. 1 except that a storage area ROM14 for storing the
bodiments, the present invention is also applicable to an weight values for constituent or elementary decisions
on/off type fire alarm system in which the decision and the weight value storage area RAM13 are to serve
concerning the fire is performed at each of the individ as a storage area RAM13 for storing the weight values
ual fire detectors, wherein only the result of the deci 55 for the overall decision or judgment. The other fire
sion is supplied to the receiving means such as the fire control panels RE' are of an identical structure to that
control panel, repeater or the like. In that case, the of the fire control panel RE shown in FIG. 1. Accord
ROM11, ROM12 and ROM13 shown as incorporated in ingly, repeated description of the these control panels
the fire control panel in FIG. 1 will be disposed in each RE' will be unnecessary. The storage area ROM14 for
of the fire detectors. Further, it is preferred that a ROM storing the weight values for the constituent or elemen
loaded with the weight values at a manufacturing stage tary decision serves to store therein for all the fire de
in a factory as mentioned above be incorporated in each tectors the weight values of the signal lines described
of the fire detectors in place of RAM12 and RAM13 in hereinafter for the purpose of obtaining the fire infor
consideration of the fact that no space is available in the nation values from each of the individual sensors incor
fire detector for providing the ten key and others shown 65 porated in each fire detector. On the other hand, the
in FIG. 1 for inputting the data in the RAM12. In that storage area RAM13 for storing the weight values for
case, the steps 401 - 408 shown in FIG. 4 would be overall decision or judgment serves to store therein for
executed by a signal processing apparatus installed at all the fire detectors the weight values provided for the
5,168,262
17 18
overall decision, as described hereinafer, for the pur element fire detectors DE1'-DEN, respectively. In the
pose of deriving the overall fire information value on net structure shown in FIG. 3A, a block A is assumed to
the basis of the individual fire information values ob be provided in association with a smoke sensor FS1, a
tained from each of the elementary or constitent sensors block B is assumed to be provided in association with a
incorporated in each of the fire detectors. temperature sensor FS2, a block C is assumed to be
Further, in the case of the multi-element fire detector provided in association with a gas sensor FS3, and a
DE', the fire phenomenon detecting means, i.e. the block D is assumed to be provided for receiving the
sensor part FS, is not of the single element structure but outputs from the blocks A-C to thereby output one
is implemented as a fire phenomenon detecting means fire probability signal on the basis of the synthetic deci
adapted for detecting a plurality of physical quantities, 10 sion or judgment of the outputs of the blocks A-C.
i.e. a multiplicity of elementary quantities such as heat, Inputted to the block A, B and C are six time-serial
smoke, gas and the like attributable to the fire phenome smoke sensor levels SLVs-SLV6, temperature sensor
non and may comprise a smoke sensor part FS which levels SLV6-SLV, and gas sensor levels
may be of a scattered light type, by way of example, a SLV-SLV6, respectively, which are collected by
temperature sensor part FS2 which may include, for 15 the fire control panel RE' from the sensor parts FS1,
example, a thermistor, a gas sensor part FS3 which may FS2 and FS3 of the associated multi-element fire detec
include, for example, a gas detecting element, together tor. In response to these inputs, the blocks A, B and C
with interfaces IF23 and IF24 provided in association output the fire probability signals OUT OUT and
with the sensor parts mentioned above. The remaining OUT respectively. These fire probability signals are
structure of the multi-element fire detector DE1" is the 20 input to the block D which then judges synthetically
same as that of the fire detector DE1 shown in FIG. 1 the input fire likelihood signals to output a more reliable
and thus description thereof will be omitted. Each of fire probability with very high accuracy.
the sensor parts FS, FS2 and FS3 include components In the case of the second embodiment of the inven
such as an amplifier, a sampling and hold circuit, an tion, it is assumed that the blocks A-C are previously
analogue-to-digital converter, etc. which are not shown 25 prepared for each fire detector already at a manufactur
in the drawings. ing stage and stored in the storage area ROM14 for the
Although the first multi-element fire detector DE1 is weight values for the element decision. According to a
shown in FIG. 1A as incorporating three sensor parts method of preparing, for example, the block. A for the
which are to serve as the fire phenomenon detecting smoke sensor, the weight values of the signal lines are
means, it should be understood that the invention is not 30 adjusted on a line-by-line basis in accordance with the
limited to the number and the types of the sensor parts expressions Eq. 1 to Eq. 6 mentioned hereinbefore with
as shown but the number and the types of the sensor the aid of the net generating program illustrated in FIG.
parts may vary from one to another multi-element fire 6 by using the definition table shown in FIGS. 2, 2" and
detector. Besides, in the case of a set in which a plurality described before in conjunction with the first embodi
of fire detectors are employed, the number and types of 35 ment of the invention. The other blocks B and C can be
fire detectors combined as a set can be altered, as occa prepared in a similar manner by adjusting the weight
sion requires. values of the relevant signal lines on a line-by-line basis
Preceding the concrete description of operation of in accordance with the expressions Eq. 1 ~Eq. 6
the second embodiment of the present invention with through the net creating rogram by preparing the defi
the aid of FIGS. 4, 5A, 6 and 7, description will first be 40 nition tables for the temperature sensor and the gas
directed to the underlying concept. sensor, respectively. In this case, the sensor level ob
With the second embodiment of the invention, it is tained by the smoke sensor part FS1 is converted to a
contemplated to collect time-serially a plurality of sen numerical value in a range of "0" to "1" which corre
sor levels from the individual sensors of plural sensor spond to a smoke concentration of 0%/m-20%/m, by
parts of the multi-element fire detector (or of plural fire 45 way of example. The sensor level obtained from the
detectors in case the multi-element fire detector is con temperature sensor part FS2 is converted into a numeri
stituted by a set of fire detectors) which are, respec cal value in a range of "0" to "1" corresponding to a
tively, adapted to detect different types of physical temperature range of 0° C. - 64 C. And, the sensor
quantities inherent to the fire phenomenon, to thereby level obtained from the gas sensor part FS3 is converted
obtain rapidly and correctly various information about 50 into a numerical value in a range of "0" to "1" which
a fire such as fire probability and the degree or level of may correspond to a concentration of carbon monoox
danger on the basis of all the sensor levels as collected. ide (CO) in a range of 0 ppm -200 ppm.
More specifically, as the plurality of time-serial sensor When the teach-in procedure of the definition table
levels, the sensor level of each sensor part is sampled such as shown in FIGS. 2, 2" for the net structure of the
every fifth second over a period of twenty-five seconds 55 blocks A-C is completed, i.e. upon completion of the
to thereby obtain the six sensor level samples in total. weight value adjustment on the line-by-line basis, these
On the basis of these sensor level samples, a fire decision weight values are then stored in the area of the storage
as to a fire is made at each of the sensor parts, being then area ROM14 assigned to the associated fire detector at
followed by the synthetic decision made on the basis of a manufacturing stage, for example, to be utilized in the
the fire information obtained from the individual sensor 60 fire monitoring operation described hereinafter.
parts, to thereby derive more reliable fire information, Next, description will be directed to the teach-in
as will be described hereinafter by reference to FIGS. 2, procedure for the net structure shown in FIG. 3A for
2, 2A, 3A, 3 and 3B. the block D. As shown in detail in FIG. 3B, the net
Before entering into description of the operation structure for the block D is so implemented that it has
outlined above, a net structure such as shown in FIG. 65 three layers at the input stage, three layers at an inter
3A will be looked at. The network structure shown in mediate stage and one layer at the output stage, where
FIG. 3A is assumed to be incorporated in the fire con nine signal lines extend between the input stage and the
trol panel RE' in a number corresponding to the multi intermediate stage while three signal lines extend be
19
5,168,262
20
tween the intermediate stage and the output stage. In gram described hereinbefore are supplied to the input
putted to input layers IN1, IN2 and IN3 are the fire stages of each of the net structures A-C, respectively,
probabilities OUT OUT, and OUT output from the in the actual fire monitoring operation, whereon the
blocks A, B and C, respectively, whereby the fire prob values OUT OUT and OUT obtained from the output
ability decided more strictly is output from the output stage OT are arithmetically determined by using the
stage OT1. corresponding weight values in accordance with the
Referring to FIG. 2A, there is shown a definition expressions Eq. 1 - Eq. 4, the values thus determined
table for teaching the net structure for the block D. being then supplied to the input stage of the net struc
Shown in three left columns of the definition table are ture D to obtain finally the fire probability OUT simi
nine combination patterns of particular values of the O larly in accordance with the expressions Eq. 1 - Eq. 4
output OUT from the net structure for the smoke sen by using the corresponding weight values.
sor part, the output OUT, from the net structure for the More specifically, referring to FIG. 4, FIG. 5A and
temperature sensor part and the output OUT from the FIG. 7, in succession to the step 409 shown in FIG. 4,
net structure for the gas sensor part, while shown at one the fire monitoring operation is performed sequentially,
right column are the accurate fire probabilities which 15 starting from the first fire detector. Describing the fire
are determined experimentally for the abovementioned monitoring operation in connection with the n-th fire
patterns, respectively. detector DE', a data send-back command is first sent
The net structure shown in FIG.3B may be prepared, out onto the signal line L through the interface IF11
for example, in the field, by adjusting the weight values from the signal transmission/reception part TRX1 to
on the basis of the contents of the definition table shown 20 the n-th fire detector DE" (step 411).
in FIG. 2A in accordance with the expressions Eq. Upon reception of the data send-back command by
1-Eq. 6 with the aid of the net creating program the n-th fire detector DE, the fire detector DE
shown in FIG. 6 in such manner as described hereinbe which is assumed to be a multi-element fire detector
fore, whereon the adjusted weight values are stored in fetches therein through the interfaces IF21, IF23 and
the storage area RAM13 for the weight values for syn 25 IF24, respectively, the sensor levels detected by the
thetic decision shown in FIG. 1 (at the step 406 in FIG. sensor parts FS1, FS2 and FS3 on the basis of the physi
4) to be utilized subsequently in the fire monitoring cal quantities such as of smoke, heat, gas and others
operation. inherent to a fire phenomenon and converted into digi
As will be appreciated from the above description, tal quantities by the incorporated analogue-to-digital
the net structure is created by teaching the definition 30 converter, wherein these sensor levels are sent back en
table. In this conjunction, it should be noted that the bloc from the signal transmission/reception part TRX2
creation of such net structure may be performed by through the interface IF22. In case the fire detecting
inputting the definition table in the fire control panel means is constituted by a set of plural fire detectors, the
RE", for example, of the fire alarm system installed in fire control panel RE' collects the sensor levels from the
the field or alternatively the weight values may be de 35 plurality of fire detectors of the set to thereby make the
termined with the aid of the net structure creating pro fire decision on the basis of the collected sensor levels.
gram at a manufacturing stage in a factory or some For the data acquisition of this kind, a conventional
other place and stored in a ROM such as an EPROM or polling technique can be adopted. It is also possible to
the like, wherein the ROM is employed in the system. In use the systems described in the specifications of the
the case of the instant embodiment, it is assumed that undermentioned patent applications 1)-3) filed in the
the weight values for the net structures of the blocks name of the same inventor and applicant as those of the
A-C are previously determined and stored in a ROM present application.
while the weight values for the net structure of the 1) In Japanese Patent Application SHO 63-168986
block D are determined in situ or in the field with the filed on Jul. 8, 1988 under the title "Fire Alarm Equip
aid of the net structure creating program. 45 ment', there is described a system in which a start ad
The above description has been made on the assump dress is assigned to a first one of fire phenomenon de
tion that the number of information values input to the tecting parts, i.e. plural sensor parts of a multi-element
input stages of the net structure A-C is six with only fire detector, while the remaining fire phenomenon
one information value output from the output stage, detecting parts are assigned with associative addresses
while in the case of the net structure D, the number of 50 associated with the start address, wherein in response to
information values input to the input stage is three with a data send-back command issued by a fire control panel
only one information value output from the output to a given one of the addresses, the fire phenomenon
stage. However, it will be readily understood that the detecting part corresponding to that address sends the
numbers of these input and output information values data as detected to the fire control panel.
can arbitrarily be selected, as occasion requires. As the 55 2) In Japanese Patent Application SHO 63-201861
information output from the output stage, there can be filed on Aug. 15, 1988 under the title "Fire Alarm
mentioned in addition to the fire probability various Equipment", there is described a system in which a
information values such as degree of danger, concentra receiving part, i.e. the fire control panel stores informa
tion or density of smoke, visibility or see-through dis tion of the type of one or a plurality of sensor parts of
tance and others. fire phenomenon detecting parts incorporated in each
When the net structures A-D shown conceptually fire detector in correspondence relation with the latter,
in FIG.3A have been prepared by storing in the storage wherein upon collection of the fire monitoring informa
area ROM14 and the RAM13 the weight values ad tion from the individual fire detectors, address signals of
justed on the line-by-line, basis through teaching of the the fire detectors to be polled are sent out together with
definition tables shown in FIGS. 2, 2 and FIG. 2A, the 65 the type information corresponding to the fire monitor
six sensor levels sampled time-serially throughout the ing information required for these type information of
period of 25 seconds for each of the sensor parts the fire detector(s), and wherein the fire detector re
FS1-FS3 through the net structure calculation pro sponds to the reception of the type information sent
5,168,262
21 22
thereto through the polling from the fire control panel At first, when the sensor levels SLV 1 - SLVs 6 origi
to thereby send out the fire monitoring information nating in the smoke sensor FS1 are input to the net
available from the fire phenomenon detecting part des structure A shown in FIG. 3A (step 514), the net struc
ignated by the abovementioned corresponding type ture calculation program 700 calcuates NET (j) in
information. accordance with the expression Eq. 1 mentioned herein
3) In Japanese Patent Application SHO 63-209356 before (step 703), the result of which is converted to the
filed on Aug. 25, 1988 under the title "Fire Alarm IMjvalue in accordance with the expression Eq. 2 (step
Equipment', there is described a system in which each 704). When the IMj values have been determined for all
of fire detectors is provided with type information of IM1-IM(J-4) (Y of step 705), then NET2(k) is calcu
fire phenomenon detecting parts incorporated in the fire O lated on the basis of the IMj values in accordance with
detector as set by first means and sends out one or a the expression Eq. 3 mentioned hereinbefore (step 708),
plurality of type information in response to a first type the result of which is converted to the value of OTk in
information request issued by a control panel, the se accordance with the expression Eq. 4 (step 709). Upon
quence of the species information as sent out being determination of OTk (k= 1 in the case of the instant
stored, wherein in response to the request for fire moni 15 embodiment), i.e. upon determination of the output
toring information from the control panel, individual OUTs of the net structure A (Y of step 710), return is
fire monitoring information obtained from one or a made to the flow chart shown in FIG. 5, whereon the
plurality of fire phenomenon detecting parts is sent out sensor levels SLV - SLV6 of the temperature sensor
in the sequence as stored, while the control panel first 20 part FS2 are then supplied to the net structure B (step
stores therein the type information received from the 515). Similarly, the output OUTt is determined by the
fire detectors in the receiving order in correspondence net structure calculation program 700 through the same
with the addresses of the fire detectors, and wherein procedure as described above, and the sensor levels
upon reception of the fire monitoring information from SLV 1 - SLV 6 from the gas sensor part FS3 are then
the fire detector, decision is made as to which of the fire supplied to the net structure C (step 516), whereby the
phenomenon detecting parts the fire monitoring infor 25 output OUTg is determined by the net structure calcu
mation as received originates by collating the receiving lation program 700.
order of the received fire monitoring information with When the outputs OUTs, OUTt and OUTg from the
the abovementioned stored type information. net structures A, B and C, respectively, have thus been
Referring back to the present invention, data sent 30 determined, these outputs are then supplied to the net
from the n-th fire detector DE', if any, (Y of step 412) structure D shown in FIG. 3B as well (step 517),
is stored in the work area RAM11 (step 413). whereon the net structure calculation program 700 is
The work area RAM11 includes areas for storing the executed in a similar manner as described above. Thus,
plurality of sensor levels for each of the fire detectors, the fire probability ratio OUT is obtained as the final
wherein the area for each of the fire detectors is so 35 output from the output stage OT of the net structure D.
segmented or partitioned that the sensor levels of the Next, the fire probabilities OUT, OUTs, OUTt and
plural constituent sensor parts sent back from the fire OUTg as obtained are displayed on the display unit DP
detector upon every polling can be stored for a prede through the interface IF12 (step 518), while the final fire
termined time. More specifically, since it is assumed in probability OUT is compared with the reference value
the case of the instant embodiment that the single pol K of the fire probability which is read out from the
ling period for the fire detectors DE1'-DEN by the various constant table storage area ROM12 (step 519).
When OUT2K, appropriate fire operations measures
fire control panel RE" is five seconds with the above such as fire display or fire alarm are taken (step 520).
mentioned predetermined time period being twenty Now, as the fire monitoring operation for the n-th fire
five seconds and that the sensor levels obtained through detector has been completed, similar fire monitoring
six pollings from each element sensor part is to be 45 operations are performed for the next fire detector.
stored, the area provided in the work area RAM11 for In the case of the embodiment of the invention de
the n-th fire detector DE' which is assumed to include scribed above, the first net structures are provided in
three element sensor parts FS1, FS2 and FS3 stores con correspondence with the plural element sensors, respec
stantly therein the sensor levels SLV 1 - SLV 6, SLV. tively, wherein the plural sensor levels collected time
1~SLV 6 and SLV 1 - SLV 6, i.e. eighteen sensor 50 serially from the individual element sensor parts are
levels in total obtained through six pollings of the three supplied to the corresponding first net structures, re
element sensor parts, respectively. In that case, the spectively, to obtain the respective fire decision infor
oldest sensor level of each element sensor part is dis mation values, which are again supplied to the second
carded every time a new sensor level is sent back upon additional net structure to thereby obtain the final fire
polling. 55 decision information value. It should, however, be ap
When the data sent back from the n-th fire detector preciated that, instead of providing the net structures in
DE", i.e. the three sensor levels from the individual correspondence with the individual element sensors,
element sensor parts, have been stored in the area for respectively, only one net structure may be provided
the n-th fire detector provided in the work area RAM11 for the whole system, wherein all the plural sensor
with the oldest data being discarded (step 413), the six levels obtained time-serially from the plurality of the
sensor levels SLV 1 - SLV 6, SLV 1 - SLV 6 and element sensor parts may be input to only one net struc
SLV 1 - SLV 6 of the individual element sensor parts ture to derive the fire decision information value based
stored in the n-th fire detector area are converted into on the synthetic judgment.
numerical values INi (i= 1 - 6) in a range of "0" to "1", Further, in place of collecting the time-serial plural
respectively, to be subsequently input to the net struc 65 sensor levels from all the constituent sensor parts, it is
tures A-C shown in FIG. 3A, whereon the execution equally possible to collect time-serially the plural sensor
of the net structure calculation program 700 shown in levels from at least one constituent sensor part while
F.G. 7 is activated. collecting only one sensor level from the remaining
5,168,262
23 24
element sensor parts, wherein these sensor levels are nificance thereof to said fire related information,
supplied to the second net structure by way of respec altering each detection information value in accor
tive first net structures or to the only one net structure dance with a weighting coefficient assigned
provided for the whole system for thereby obtaining the thereto, and arithmetically determining said fire
fire decision information. 5 related information value on the basis of the thus
Although it has been assumed in the foregoing that altered detection information values.
the plurality of fire phenomenon detecting means are of 2. A fire alarm system for signal processing detection
mutually different types, it should be appreciated that information values generated by a plurality of fire phe
the plurality of fire phenomenon detecting means may nomenon detecting devices to obtain a value denoting at
be of a same type installed at different locations (in a O least one type of fire related information, said fire alarm
same room or zone). In that case, the definition table system comprising:
shown in FIG. 2A is so prepared that various fire deci detection information collecting means for collecting
sion values are derived from the outputs of the same detection information values from each of said fire
type sensor parts. phenomenon detecting devices, wherein a plurality
It should be added that in the case of the embodiment 15 of detection information values from at least one of
of the invention, the net structures of the blocks A-C said fire phenomenon detection devices is time-seri
shown in FIG. 3A are created at a manufacturing stage ally collected; and
in a factory and the weight values for the net structures signal processing means for signal processing said
are stored in the weight value storage ROM14 for the detection information values collected by said de
element decision such as an EPROM or the like, while 20 tecting information collecting means by respec
only the net structure for the block D shown in FIG. tively assigning weighting coefficients to each
3A is generated by executing the net structure creating input detection information value in accordance
program with the weight value for the overall or syn with a relative significance thereof to said fire re
thetic decision being stored in the storage area RAM13. lated information, altering each detection informa
It should, however, be appreciated that the weight 25 tion value in accordance with a weighting coeffici
values of all the net structures for all the blocks A-D ent assigned thereto, and arithmetically determin
may be stored in the storage RAM13 through the net ing said fire related information values on the basis
structure creating program after the installation of the of the thus altered detection information values.
fire detector or reversely all the net structures may be 3. A fire alarm system as set forth in claim 2, wherein
previously created in manufacturing steps in a factory 30 said signal processing means includes first auxiliary
for allowing a ROM such as an EPROM storing the processing means associated with said at least one fire
weight values for these net structures to be employed, phenomenon detecting device by which said plural
as will readily be apparent for those skilled in the art. detection information values are time-serially collected,
Besides, in lieu of the analogue type fire alarm system for performing an arithmetic operation to obtain indi
described above in conjunction with the exemplary 35 vidual fire information values, and second auxiliary
embodiments, the present invention is also applicable to processing means for processing the individual fire in
an on/off type fire alarm system in which the decision formation values input from said first auxiliary process
concerning a fire is performed on the side of the individ ing means and the detection information values input
ual fire detectors, wherein only the result of the deci from the remaining fire phenomenon detecting devices
sion is supplied to the receiving means such as a fire in which detection information values are not time-seri
control panel, repeater or the like. In that case, the ally collected, to thereby derive final fire related infor
ROM11 and ROM12 shown as incorporated in the fire mation values of higher reliability.
control panel in FIG. 1A are disposed in each of the fire 4. A fire alarm system for obtaining a value denoting
detectors. Further, it is preferred that a ROM loaded at least one type of fire related information by process
with the weight values at a manufacturing stage in a 45 ing signals representative of detection information gen
factory as mentioned above is incorporated in each of erated by a plurality of fire phenomenon detecting de
the fire detectors in place of the ROM14, RAM12 and vices, comprising:
RAM13 in consideration of the fact that no room is detection information collecting means for time-seri
available in the fire detector for providing the ten key ally collecting a plurality of detection information
and others shown in FIG. 1 or FIG. 1A for inputting 50 values from each of said fire phenomenon detecting
the data in the RAM12. devices; and
I claim: signal processing means for signal processing the
1. A fire alarm system for receiving detection infor detection information values collected by said de
mation values generated by at least one fire phenome tection information collecting means, and includ
non detecting device and for subjecting the detection 55 ing means for respectively assigning weighting
information values to signal processing for obtaining a coefficients to each of said detection information
value denoting at least one type of fire related informa values upon inputting thereof in accordance with
tion, said fire alarm system comprising: an extent to which each input detection informa
detection information collecting means for time-seri tion value contributes to said related fire informa
ally collecting a plurality of detection information tion, altering each detection information value in
values generated by said at least one fire phenome accordance with a weighting coefficient assigned
non detecting device; and thereto, and arithmetically determining said fire
signal processing means for signal processing the related information value on the basis of the thus
plurality of the detection information values time altered detection information values.
serially collected by said detecting information 65 5. A fire alarm system as set forth in claim 4, wherein
collecting means by respectively assigning said signal processing means includes first auxiliary
weighting coefficients to each input detection in processing means associated with said fire phenomenon
formation value in accordance with a relative sig detecting devices in which said plural detection infor
25
5,168,262
26
mation values are time-serially collected, for perform cessing means when said particular set of detection
ing an arithmetic operation to obtain individual fire information values stored in said table is input approxi
information values, and second auxiliary processing mates said fire related information value stored in said
means for processing the individual fire information table, wherein said signal processing means includes
value input from said first auxiliary processing means to storage means for storing weighting coefficients for the
derive final fire related information values of higher respective detection information values, said weighting
reliability. coefficients stored in said storage means for said first
6. A fire alarm system as set forth in any one of claims auxiliary processing means being previously established
1 to 5, wherein said signal processing means includes so that the fire related information value arithmetically
storage means for storing in advance the weighting 10 determined by said first auxiliary processing means
coefficients assigned to the detection information val when the particularly set of detection information val
ues, respectively, said weighting coefficients being se ues for said first auxiliary processing means is input
lected to cause the fire related information value arith approximates a desired information value which is to be
metically determined by said signal processing means in derived from each particular set for said first auxiliary
response to the inputting of a particular set of the detec 15 processing means, while the weighting coefficients
tion information values to approximate a desired fire stored in said storage area for said second auxiliary
related information value which is to be derived from processing means are initially adjusted by said adjusting
said particular set of the information values. means on the basis of the contents of said table.
7. A fire alarm system as set forth in any one of claims 9. A fire alarm system as set forth in any one of claims
1 to 5, further comprising a table having stored therein 20 1 to 5, comprising a receiving part and a plurality of fire
a particular set of detection information values together detectors connected to said receiving part, each of said
with at least one fire related information value which is fire detectors including at least one fire phenomenon
to be obtained when said particularly set of detection detecting means for detecting a physical quantity attrib
information values is input, adjusting means for adjust utable to a fire phenomenon, wherein said signal pro
ing the weighting coefficients so that said fire related 25 cessing means is incorporated in said receiving part.
information value arithmetically determined by said 10. A fire alarm system as set forth in any one of
signal processing means when said particularly set of claims 1 to 5, comprising a receiving part and a plurality
detection information values stored in said table is input of fire detectors connected to said receiving part, each
approximates said fire related information values stored of said fire detectors including at least one fire phenom
in said table, wherein said signal processing means in 30 enon detecting means for detecting a physical quantity
cludes storage means for storing weighting coefficients attributable to a fire phenomenon, wherein said signal
for said respective detection information values, said processing means is incorporated in each of said fire
weighting coefficients stored in said storage area being detectors.
first adjusted by said adjusting means on the basis of the 11. A fire alarm system as set forth in claim 3 or 5,
contents of said table. 35 comprising a receiving part and a plurality of fire detec
8. A fire alarm system as set forth in claim 3 or 5, tors connected to said fire receiving part, each of said
further comprising a table for storing therein a particu fire detectors including at least one fire phenomenon
lar set of detection information values together with at detecting means for detecting a physical quantity attrib
least one fire related information value which is to be utable to a fire phenomenon, wherein said first auxiliary
obtained when said particular set of detection informa processing means is incorporated in each of said fire
tion values is input, adjusting means for adjusting the detectors and said second auxiliary processing means
weighting coefficients so that said fire related informa being incorporated in said receiving part.
tion value arithmetically determined by said signal pro x 2k s
45
50
55
65

Assess Now ESE A: sEF 2 E

Uploaded by

Copyright:

Available Formats

Assess Now ESE A: sEF 2 E

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Assess Now ESE A: sEF 2 E

Uploaded by

Copyright:

Available Formats

||||||||III US005 68262A

United States Patent (19) 11 Patent Number: 5,168,262

- as a re- - -a rere rere- a-- re- - -- a re-3- re-a- - - -

PROGRAM ROMZEERAMa ROM Ele

assess now ESE a

INPUT 0.90 0.380 0.320 0.380 0.260

SEND DATA SEND-BACK - 41

INPUT SLV-NSLV6 TO 515

INPUT SLVg NSLVg6 TO 516

F. G. 6 NET STRUCTURE 600

ALL WEIGHT VAL OF SIG

INTERMEDIATE STAGE - OUTPUT STAGE

You might also like