CET205 M3 Ktunotes - in
CET205 M3 Ktunotes - in
CET205 M3 Ktunotes - in
TRAVERSE SURVEYING
Traversing is that type of survey in which a number of connected survey lines from the
framework and the directions and lengths of the survey lines are measured with the help of an
angle (or direction) measuring instrument and a tape (or chain) respectively. When the lines from
a circuit which ends at starting point, it is known as a closed traverse. If the circuit ends
elsewhere, it is said to be an open traverse. The closed traverse is suitable for locating the
boundaries of lakes, woods etc., and for the survey of large areas. The open traverse is suitable
for surveying a long narrow strip of land as required for a road or canal or the coast line.
METHODS OF TRAVERSING
I. CHAIN TRAVERSING
In this method, the whole of the work is done with the chain and tape. No angle
measuring instrument is used and the directions of the lines are fixed entirely by liner
measurements. Angles fixed by linear or tie measurements are known as chain angles.
Fig. (a) Shows a closed chain traverse. At A, the directions AB and AD are fixed by internal
measurements Aa1, Ad1 and a1d1. However, the direction may also be fixed by external
measurements such as at station B. fig. (b) Shows an open chain traverse. The method is
unsuitable for accurate work and is generally not used if an angle measuring instrument such as a
compass, sextant or theodolite is available.
c2
b2 B c1 C
b1
A
(a)
C
C2
B c1
b2
b1
(b)
In chain and compass traversing, the magnetic bearings of the survey lines are measured
by a compass and the lengths of the lines are measured either with a chain or with a tape. The
direction of magnetic meridian is established at each traverse station independently. The method
is also known as free or loose needle method. A theodolite fitted with a compass may also be
used for measuring the magnetic bearings of the traverse line. However, the method is not so
accurate as that of transit tape traversing. The methods of taking the details are almost the same
as for chain surveying.
In this method also, the magnetic bearings of traverse lines are measured by a theodolite
fitted with a compass. However, the direction of the magnetic meridian is not established at each
station but instead, the magnetic bearings of the lines are measured with reference so the
direction of magnetic meridian established at the first station. The method is therefore more
accurate than the loose needle method. The lengths of the lines are measured with a 20 m or 30
m steel tape. There are three methods of observing the bearing of lines by fast needle method.
i. Direct method with transiting
ii. Direct method without transiting
iii. Back bearing method
Set the theodolite at P and level it. Set the vernier A exactly to zero reading. Loose the
clamp of the magnetic needle. Using lower clamp and tangent screw, point the telescope
to magnetic meridian.
Loose the upper clamp and rotate the telescope clockwise to sight Q. bisect Q accurately
by using upper tangent screw. Read vernier A which gives magnetic bearing of the line
PQ.
P R
With both the clamps clamped, move the instrument and set up at Q. using lower clamp
and tangent screw, take a book sight on P. See that the reading on the vernier A is still the
same as the bearing of PQ.
Transit the telescope. The line of sight will now be in the direction of PQ while the
instrument reads bearing of PQ. The instrument is therefore oriented.
Using the upper clamp and tangent screw, taken a foresight on R. Read vernier A which
gives the magnetic bearing of QR.
Continue the process at other stations. It is to be noted here that the telescope will be
normal at one station and inverted at the next station. The method is therefore suitable
only if instrument is in adjustment.
Set the instrument at P and orient the line of sight in magnetic meridian.
Using upper clamp and tangent screw tale a foresight on Q. the reading on vernier A
gives the magnetic bearing of PQ.
P R
With both plates clamped, move the instrument and set it at Q. Take a backsight on P.
check the reading on vernier A which should the same as before. The line of sight is out
of orientation by 1800.
Loosen the upper clamp and rotate the instrument clockwise to take a foresight on R.
Read vernier. Since the orientation at Q is 1800 out, a correction of 1800 is to be applied
to the vernier reading to get the correct bearing of QR. Add 1800 if the reading on the
vernier is less than 1800 and subtract 1800 if it is more than 1800.
Shift the instrument of R and take backsight on Q. The orientation at R will be out by
1800 with respect to that at Q and 3600 with respect to that at P. Thus, after taking a
foresight on the next station, the vernier reading will directly give magnetic bearing of
the line, without applying any correction of 1800.
The application of 1800 correction is therefore necessary only at 2nd, 4th, 6th station occupied.
Instead of applying correction at even station, opposite vernier may be read alternatively, i.e.,
vernier A at P, vernier B at Q, vernier A at R, etc. However, it is always convenient to read one
vernier throughout and apply the correction at alternate stations.
Of the three methods of fast needle, the second method is the most satisfactory.
In this method, the angles between the lines are directly measured by a theodolite. The
method is therefore most accurate in comparison to the previous three methods. The magnetic
bearing of any line can also be measured (if required) and the magnetic bearing of other lines can
be calculated. The angles measured at different stations may be either (a) included angles or (b)
deflection angles.
An included angle at a station is either of the two angles formed by the two survey lines
meeting there. The method consists simply in measuring each angle directly from a backsight on
the preceding station. The angles may also be measured by repetition, if so desired. Both face
observations must be taken and both the vernier should be read. Included angles can be measured
either clockwise or counter-clockwise but it is better to measure all angles clockwise, since the
graduations of the theodolite circle increase in this direction. The angles measured clockwise
from the back station may be interior or exterior depending upon the direction of progress round
the survey. Thus, in Fig (a), direction of progress is counter-clockwise and hence the angles
measured clockwise are directly the interior angles. In Fig (b), the direction of progress around
the survey is clockwise and hence angles measured clockwise are exterior angles.
Q R Q R
P S P S
(a) (b)
A deflection angle is the angle which a survey line makes with the prolongation of the
preceding line. It is designed as right (R) or left (L) according as it is measured clockwise or
anti-clockwise from the prolongation of the previous line.
This method of traversing is more suitable for survey of roads, railways, pipe-lines etc.
where the survey lines make small deflection angles. Great care must be taken in recording and
plotting whether it is right deflection angle or left deflection angle. However, except for
specialized work in which deflection angles are required, it is preferable to read the included
angles by reading clockwise from the back station. Lengths of lines are measured precisely using
a steel tape.
Plane table traverse involves the same principles s a transit traverse. At each
successive station the table is set, a foresight is taken to the following station and its location is
plotted by measuring the distance between the two stations as in the radiation method described
earlier. Hence, traversing is not much different from radiation as far as working principles are
concerned --- the only difference is that in the case of radiation the observations are taken to
those points which are to be detailed or mapped while in the case of traversing the observations
are made to those points which will subsequently be used as instrument stations. The method is
widely used to lay down survey lines between the instrument stations of a closed or unclosed
traverse.
PROCEDURE
Set the table at . Use plumbing fork for transferring on to the sheet. Draw the
direction of magnetic meridian with the help of trough compass.
With the alidade pivoted about a, sight it to and draw the ray. Measure and
scale off to some scale. Similarly, draw ray towards measure and plot .
Shift the table to and set it. Orient the table accurately by back-sighting . clamp
the table.
Pivoting the alidade about , sight to . measure and plot it on the drawn ray to
the same scale. Similarly, the table can be set at other stations and the traverse is
completed.
d
c d
c
e e
b b
a a
check D C
check
E
e
e
b b
a a
A B
A point can be located from a transit station by taking an angle to the point and measuring
the corresponding distance from the station to the point. Any numbers of points can thus be
located. The angles are usually taken from the same backsight, as shown in Fig. The method is
suitable specially when the details are near the transit station.
To C
If the point or points are away from the transit stations or if linear measurements cannot
be made, the point can be located by measuring angles to the point from at least two stations.
This method is also known as method of intersection. For good intersection, the angle to point
should not be less than 200.
A B
To C
(3) Locating by distances from two stations
Fig. illustrates the method of locating a point by measuring angle at one station and
distance from the other. The method is suitable when point is inaccessible from station at which
angle is measured.
distance
A B
If the point is near a transit line but is away from the transit station, it can located by
measuring its distance from two points on traverse line. The method is more suitable if such
reference points (such as x and y in Fig.) are full chain points so that can be staked when the
traverse line is being chained.
Distance Distance
From A x y To B
If points to be detailed are more and are near to traverse line, they can be located by
taking offsets to the points as explained in chain surveying. The offsets may be oblique or may
be perpendicular.
The errors involved in traversing are two kinds: linear and angular. For important work
the most satisfactory method of checking the linear measurements consists in chaining each
survey line a second time, preferably in reverse direction on different dates and by different
parties. The following are the checks for the angular work:
a) sum of measured interior angles should be equal to (2N---4) right angles, where
N=numbers of sides of the traverse.
b) If the exterior angles are measured, their sum should be equal to (2N + 4) right angles.
The algebraic sum of deflection angles should equal to 3600, taking the right-hand deflection
angles as positive and left-hand angles as negative.
The fore bearing of the last line should be equal to its back bearing ± 1800 measured at the
initial station.
D E E
P. D
C F
A C Q
A B G
(a) (b)
TRAVERSE COMPUTATIONS
The latitude of a survey line may be defined as its co-ordinate length measured
parallel to an assumed meridian direction (i.e. true north or magnetic north or any other reference
direction). The departure of survey line may be defined as its co-ordinate length measured at
right angles to the meridian direction. The latitude (L) of the line is positive when measured
northward (or upward) and is termed as northing; the latitude is negative when measured
southward (or downward) and is termed as southing. Similarly, the departure (D) of the line is
positive when measured eastward and is termed as easting, the departure is negative when
measured westward and is termed as westing.
D=l sin
B
IV I
(+, -) l (+, +)
L=l cos
III II
(-, -) (-, +)
Thus in Fig. the latitude and departure of the line AB of length 1 and reduced bearing are
given by
L = + 1 cos
D = + 1 sin
To calculate the latitudes and departure of the traverse lines, therefore, it is first essential to
reduce the bearing in the quadrantal system. The sign of latitudes and departures will depend
upon the reduced bearing of a line. The following table gives signs of latitudes and departures.
Thus, latitude and departure co-ordinates of any point with reference to the preceding point
are equal to the latitude and departure of the line joining the preceding point to the point under
consideration. Such co-ordinates are also known as consecutive co-ordinates or dependent co-
ordinates.
INDEPENDENT CO-ORDINATE
The co-ordinates of traverse stations can be calculated with respect to a common origin.
The total latitude and departure of any point with respect to a common origin are known as
independent co-ordinates or total co-ordinates of the point. The two reference axes in this case
may be chosen to pass through any of the traverse station but generally a most westerly station is
chosen for this purpose. The independent co-ordinates of any point may be obtained by adding
algebraically the latitudes and the departure of the lines between that point and the origin.
Thus, total latitude (or departure) of end point of a traverse = total latitudes (or departures).
The axes are so chosen that the whole of the survey lines lie in the north east quadrant with
respect to the origin so that the co-ordinates of all the points are positive. To achieve this,
arbitrary values of co-ordinates are assigned to the starting point and co-ordinates of other points
are calculated.
CLOSING ERROR
If a closed traverse is plotted according to the field measurements, the end point of the
traverse will coincide exactly with the starting point, owing to the errors in the field
measurements of angles and distances. Such error is known as closing error.
C D
B
Closing error A`
In a closed traverse, algebraic sum of the latitudes (i.e. ) should be zero and the
algebraic sum of the departures (i.e. ) should be zero. The error of closure for such traverse
may ascertained by finding and , both of these being the components of error parallel
and perpendicular to the meridian.
tan =
The sign of and will thus define the quadrant in which the closing error lies. The
relative error of closure is
= = = ⁄
Before calculating latitudes and departures, the traverse angles should be adjusted to
satisfy geometric conditions. In a closed traverse, the sum of interior angles should be equal to
(2N – 4) right angles (or the algebraic sum of deflection angles should be 3600). If the angles are
measured with the same degree of precision, the error in the sum of angles may be distributed
equally to each angle of the traverse. If the angular error is small, it may be arbitrarily distributed
among two or three angles.
Adjustment of bearings
in a closed traverse in which bearings are observed, the closing error in bearing may be
determined by comparing the two bearings of the last line as observed at the first and last stations
of traverse. Let be the closing error in bearing of last line of a closed traverse having N sides.
We get
1.Bowditch’s method
The basis of this method is on the assumptions that the errors in linear measurements are
proportional to and that the errors in angular measurements are inversely proportional to
where l is the length of a line. The Bowditch’s rule, also termed as the compass rule, is mostly
used to balance a traverse where linear and angular measurements are of equal precision. The
total error in latitude and in the departure is distributed in proportion to the lengths of the sides.
Thus if,
CL = correction to latitude of any side
CD = correction of departure of any side
= total error in latitude
= total error in departure
= length of the perimeter
l = length of any side
We have, CL = x and CD = x
2.Transit method
The transit rule may be employed where angular measurements are more precise than the
linear measurements. According to this rule, the total error in latitudes and in departures is
distributed in proportion to the latitudes and departures of the sides. It is claimed that the angles
are less affected by corrections applied by transit method than by those by Bowditich’s method.
x
Thus if,
L = latitude of any line
D = departure of any line
L1 = arithmetic sum of latitudes
D1 = arithmetic sum of departure
We have, CL = x and CD = x
3.Graphical method
For rough survey, such as a compass traverse, the Bowditch rule may be applied
graphically without doing theoretical calculations. Thus, according to the graphical method, it is
not necessary to calculate latitude and departures etc. however, before plotting traverse directly
from the field notes the angles or bearings may be adjusted to satisfy geometric conditions of the
traverse.
D` C`
E` D C
E
a
d e
B` b c
A`
A B A` B` C` D` E` A`
(a) (b)
Traverse computations are usually done in a tabular form, a more common form being
Gales traverse table. For complete traverse computations, the following steps are usually
necessary:
right angles and exterior angles (2N + 4) right in the case of a compass traverse, the
bearings are adjusted for local attraction, if any.
Starting with observed bearings of one line, calculate the bearings of all other lines.
Reduce all bearings to quadrantal system.
Calculate the consecutive co-ordinates (i.e latitudes and departures).
Calculate and
Apply necessary corrections to the latitudes and departures of the lines so that =0
and = 0. The corrections may be applied either by transit rule or by compass rule
depending upon the type of traverse.
Using the corrected consecutive co-ordinates, calculate the independent co-ordinates to
the points so that they are all positive, the whole of the traverse thus lying in the North
east quadrant.
Problem 3: The length and bearing of a closed traverse ABCD, as observed with transit
theodolite, are given below. Prepare a Gale’s traverse table.
Geometric check:
Sum of interior angle = (2N – 4)900
N=4 = 3600
correction = -24`
Bearing of any line = [bearing of previous line + measured clockwise angle] ± 1800
= 900
= 35806`
= 266042`
WCB:
FAB = FBA ± 1800 = 266042` – 1800 = 86042`
Hence check is ok
FBC = FCB ± 1800 = 35806` – 1800 = 17806`
FCD = FDC ± 1800 = 900 + 1800 = 2700
FDA = FAD ± 1800 = 1820 – 1800 = 20
RB:
AB – N 86042` E
BC – S 1054`E
CD – S 900 W
DA – N 20 E
CL = x
(CL)AB = = 0.036
(CL)BC = = 0.306
(CL)CD = =0
(CL)DA = = 0.268
CD = x
(CD)AB = = 0.7
(CD)BC = = 0.011
(CD)CD = = 0.718
(CD)DA = = 0.011
Statio Line Included Corrected RB Consecutive coordinates Correction Corrected consecutive coordinates Independent
n And angle included coordinates
length angle
L= cos D= sin L D L D L D
N S E W N S E W N S E W N E
SURVEYING & GEOMATICS
A 950241 950181 107.93 3.77 +0.268 -0.011 108.198 3.76 300 400
assumption
AB N860421E
250
B 880421 880361 14.39 249.59 +0.036 -0.7 14.43 248.89 300 + 14.43
=
314.43
BC S10541E
123
CD N900W
256
DA N20E
108
Problem 4: The length and bearing of a closed traverse ABCDE, as observed with transit
theodolite, are given below. Prepare a Gale’s traverse table.
Closing error =
angles bearing bearing
Latitude Departure
point
Correction
Inst station
Northing Southing Westing Easting Northing Southing Westing Easting
+
28705411011 7205115011 N.W.1V AB
SURVEYING & GEOMATICS
69.90
m
Total 540011011 11 540001011 Total 133.983 134.955 70.353 70.683 5.5182 0.4536 0.1876 0.1438
0.972 0.331
Module 3
SURVEYING & GEOMATICS Module 3
OMITTED MEASUREMENTS
There are two principal methods of plotting a traverse survey: (1) the angle and distance
method, and (2) the co-ordinate method. If the length and bearing of a survey line are known, it
can be representation on plan by two rectangular co-ordinates. The axes of the co-ordinates are
the north & south line, and the east & west line. The latitude of survey line may be defined as its
co-ordinate length measured parallel to the meridian direction. The departure of the survey line
may be defined as its co-ordinate length measured at right angles to the meridian direction. The
latitude (L) of the line is positive when measured northward (or upward) and is termed as
northing. The latitude is negative when measured southward (or downward) and is termed as
southing. Similarly, the departure (D) of the line is positive when measured eastward and is
termed as easting. The departure is negative when measured westward and is termed as westing.
N
D1(+)
A
D4(-) l1 L1(+)
D
L4(+) l4
W O E
L3(-)
l3
C D3(-) l2 L2(-)
B
D2(+)
Thus, in Fig., the latitude and departure of the line OA of length l1 and reduced bearing is
given by
L1 = + l1 cos 1
&
D1 = + l1 sin
To calculate the latitudes and departures of the traverse lines, therefore, it is first essential to
reduce the bearing in the quadrantal system. The sign of latitude and departures will depend upon
the reduced bearing of line.
The following table gives the signs of latitudes and departures.
Thus, latitude and departure co-ordinates of any point with reference to the preceding point are
equal to the latitude and departure of the line joining the preceding point to the point under
consideration. Such co-ordinates are also known as consecutive co-ordinates or dependent co-
ordinates.
Independent co-ordinates
The co-ordinates of traverse station can be calculated with respect to a common origin. The
total latitude and departure of any point with respect to a common origin are known as
independent co-ordinates or total co-ordinates of the point. The two reference axes in this case
may be chosen to pass through any of the traverse stations but generally a most westerly station
is chosen for this purpose. The independent co-ordinates of any point may be obtained by adding
algebraically the latitudes and the departure of the lies between the point and the origin.
Thus, total latitude (or departure) of end point of a traverse = total latitudes (or departures)
of first point of traverse plus the algebraic sum of all the latitudes (or departures).
OMITTED MEASUREMENTS
In order to have a check on field work and in order to balance a traverse, the length and
direction of each line is generally measured in the field. There are times, however, when it is not
possible to take all measurements due to obstacles or because of some over-sight. Such omitted
measurements or missing quantities can be calculated by latitudes and departures provided the
quantities required are not more than two. In such cases, there can be no check on the field work
nor can the survey be balanced. All errors propagated throughout the survey are thrown into the
computed values of the missing quantities.
Where l1, l2, l3……etc, are the lengths of the lines and , , , …..etc. their reduced
bearings. With the help of the above two equations, the two missing quantities can be calculated.
Table 8.2 below gives the trigonometric relations of a line with its latitude and departure, and
may be used for the computation of omitted measurements.
In case (I), only one side is affected. In case II, III and IV two sides are affected both of
which may either be adjacent or may be away.
In Fig. 8.2, let it be required to calculate either bearing or length or both bearing and length
of the line . Calculate and of the four known sides and .
4 3
E C
5 2
A 1 B
Then
= Latitude of EA + =0
Or
Latitude of EA = -
Similarly, = Departure of EA + =0
Or
Departure EA = -
Knowing latitude and departure of EA, its length and bearing can be calculated by proper
trigonometrical relations.
CASE II: LENGTH OF ONE SIDE AND BEARING OF ANOTHER SIDE OMITTED
In Fig. 8.3, let the length of and bearing of be omitted. Join which becomes
the closing line of the traverse in which all the quantities are known. Thus the length and
bearing of can be calculated as in case I.
D
3
4 C
E closing line
5 2
A 1 B
In the length of sides and are known, and angle ( ) is known. The
angle and the length can be calculated as under.
Sin = sin
= 1800 – ( – )
DE = EA = DA
Example 5 : The table below gives the lengths and bearings of the lines of a traverse
the length and bearing of EA having been omitted. Calculated the length and bearing of the line
.
Solution. Fig. 8.2 shows the traverse in which is the closing line of the polygon.
Knowing the length and bearing of the lines and , their latitudes and departures
can be calculated and tabulated as under:
tan = = or = 00 28`
Example 6: A closed traverse was conducted round an obstacle and the following observations
were made. Work out the missing quantities:
Solution. The affected sides are adjacent. Fig. 8.3 shows the traverse in which is
closing line of the polygon . The latitude and departure of the closing line can be
calculated. The calculations are shown in the tabular form below:
tan = = = 300 4`
THEORY OF ERRORS
TYPES OF ERRORS
Errors of measurement are of three kinds: (i) mistakes (ii) systematic errors, and (iii) accidental
error.
(i) Mistakes
Mistakes are errors that arise from inattention, in experience, carelessness and poor
judgement or confusion in the mind of the observer.
DEFINITIONS
Conditioned Quantity. A conditioned quantity is the one whose value is dependent upon the
values of one or more quantities. Its value bears a rigid relationship to some other quantity or
quantities. It is also called a dependent quantity . For example, in a triangle ABC, <A + < B+ <C
= 1800. In this conditioned equation any two angles may be regarded as independent and the
third as dependent or conditioned.
Direct Observation. An observation is the numerical value of a measured quantity, and may be
either direct or indirect. A direct observation is the one made directly on the quantity being
determined, e.g., the measurement of a base, the single measurement of an angle etc.
Indirect Observation. An indirect observation is one in which the observed value is deducted
from the measurement of some related quantities, e.g., the measurement of angle by repetition (a
multiple of the angle being measured).
Observed Value of a Quantity. An observed value of a quantity is the value obtained when it is
corrected for all the known errors.
True Value of Quantity. The true value of a quantity is the value which is absolutely free from
all the errors. The true value of a quantity is indeterminate since the true error is never known.
Most Probable Value. The most probable value of a quantity is the one which has more chances
of being true than has any other. It is deduced from the several measurements on which it is
based.
True Error. A true error is the difference between the true value of a quantity and its observed
value.
Most Probable Error. The most probable error is defined as that quantity which added to, and
subtracted from, the most probable value fixes the limits within which it is an even chance the
true value of the measured quantity must lie.
Residual Error. A residual error is the difference between the most probable value of a quantity
and its observed value.
Observation Equation. An observation equation is the relation between the observed quantity
and its numerical value.
Conditioned Equation. A conditioned equation is the equation expressing the relation existing
between the several dependent quantities.
Normal Equation. A normal equation is the one which is formed by multiplying each equation
by the co-efficient of the unknown whose normal equation is to be found and by adding the
equations thus formed. As the number of normal equations is the same as the number of
unknowns. the most probable values of the unknown can be found from these equations.
Investigations of observations of various types show that accidental errors follow a definite law,
the law of probability. This law defines the occurrence of errors and can be expressed in the form
of equation which is used to compute the probable value or the probable precision of a quantity.
The most important features of accidental errors which usually occur are :
1. Small errors tend to be more frequent than the large ones; that is they are the most
probable.
2. Positive and negative errors of the same size happen with equal frequency; that is, they
are equally probable.
Probability Curve. The theory of probability describes these features by stating that the relative
frequencies of errors of different extents can be represented by a curve as shown in figure.
Probability Curve
This curve, called the curve of error or probability curve, forms the basis for the mathematical
derivation of theory of errors.
The formula for probable error is difficult to derive. It is stated here categorically:
Probable error of a single measurement is given by
Es = 0.6745
v 2
n 1
where Es = Probable error of single observation.
v = Difference between any single observation and the mean (or average) of the series.
n = Number of observations in the series.
Probable Error of an Average. Since the average of n measurements is the sum of the
n
measurements divided by n, the probable error of the average of n measurements is times the
n
probable error of one measurement. Thus, probable error of an average or mean is given by
Em =
n
0.6745
v = 0.6745 v = Es
2 2
n n 1 n(n 1) n
where Em = probable error of the mean.
Probable Error of a Sum. When a measurement is the result of the sums and differences of
several (n) observations having different probable errors E1, E2, E3….En, the probable error of
the measurement is the square root of the sum of the squares of the probable errors of the several
observations. Thus,
E E E .......... E n
2 2 2 2
Probable error of measurement = 1 2 3
Most Probable Value. As defined earlier the most probable value of a quantity is the one which
has more chances of being true than any other. It can be proved from the theory of errors that:
i. The most probable value of a quantity is equal to the arithmetic mean if the observations
are of equal weight.
ii. The most probable value of a quantity is equal to the weighted arithmetic mean in case
of observations of unequal weights.
Average Error. An error in a series of observation of equal weight is defined as the arithmetical
mean of separate errors, taken all with the same sign, either plus or minus.
Mean Square Error (m.s.e.). The mean square error is equal to the square root of the arithmetic
mean of the squares of the individual errors.
2
v1 v2 v3 ......... v
2 2 2
Thus, m.s.e. =
n
n
GENERAL PRINCIPLES OF LEAST SQUARES
According to the principle of least squares, the most probable value of an observed quantity
available from a given set of observations is the one for which the sum of the squares of the
residual errors is a minimum.
The arithmetic mean is the true value where the number of observed value is very large.
Hence the sum of the residuals equals zero and the sum of plus residual equals the sum of the
minus residuals.
The sum of the squares of the residuals found by the use of the arithmetic mean is a minimum.
This is thus, the fundamental law of least squares.
LAWS OF WEIGHTS
From the method of least squares the following laws of weights are established:
(1) The weight of the arithmetic mean of the measurements of unit weight is equal to
the number of observations.
For example, let an angle A be measured six times, the following being the values:
<A Weight <A Weight
1
Arithmetic mean = 300 20’ + (8” + 10” + 7” + 10” + 9” + 10”) = 300 20’ 9”
6
Weight of arithmetic mean = number of observations = 6
(2) The weight of the weighted arithmetic mean is equal to the sum of the individual
weights.
1 1 4
Weight of α + β (=720 50’ 30”) =
1 1 3 3
4 2 4
1 1 4
Weight of α - β (=110 30’ 10”) =
1 1 3 3
4 2 4
(4) If a quantity of given weight is multiplied by a factor, the weight of the result obtained
by dividing its given weight by the square of the factor.
(5) If a quantity of given weight is divided by a factor, the weight of the result is
Es = 0.6745
v 2
n 1
where v v1 v2 v3 .... vn ;
2 2 2 2 2
v = residual error
n = number of observations.
Em = 0.6745 v 2
=
Es
n(n 1) n
Case 2. Direct Observations of Unequal Weights on a Single Quantity
From the principle of least squares, the most probable values of the observed quantities (of
unequal weights or precision) are those that render the sum of the weighted squares of the
residual errors a minimum.
The most probable value of the observed quantity is equal to the weighted arithmetic mean of the
observed quantities.
= Es = 0.6745
wv 2
n 1
E 0.6745
s wv 2
w w(n 1)
(c) Probable error of weighted arithmetic mean
0.6745
wv 2
w *(n 1)
Case 3. Probable Error of Computed Quantities
The probable error of computed quantities follow the following laws depending upon the relation
between the computed quantity and the observed quantity.
1. If a computed quantity is equal to sum or difference of the observed quantity plus or minus
a constant, the probable error of the computed quantity is the same as that of the observed
quantity.
3. If a computed quantity is equal to the sum of two or more observed quantities, the p.e. of the
computed quantity is equal to the square root of the sum of the square of p.e.’s of observed
quantities.
5. If a computed quantity is a function of two more observed quantities, its probable error is
equal to the square root of summation of the squares of the p.e. of the observed quantity
multiplied by its differentiation with respect to that quantity.
ea=
= = = 1.19
0.6745
wv 2
0.6745
4
0.95.
n 1 3 1
(b) p.e. of weighted arithmetic mean
0.6745
wv 0.6745
2
4
0.36
w(n 1) 7*2
Whenever observations are made in the field, it is always necessary to check for the closing
error, if any, the closing error should be distributed to the observed quantities. For example, the
sum of the angles measured at a central angle should be 3600; if the sum is not equal to 3600, the
error should be distributed to the observed angles after giving proper weightage to the
observations. The following rules should be applied for distribution of errors:
Example 2: The following are the three angles observed at a station P closing the
horizon, along with their probable errors of measurement. Determine their correct values.
780 12’ 12” 1360 48’ 30” 1440 59’ 08”
Solution:
Example 3. An angle A was measured by different persons and the following are the values
Solution:
The most probable value of an angle is equal to its weighted arithmetic mean.
650 30’ 10” * 2 = 1310 00’ 20”
650 29’ 50” * 3 = 1960 29’ 30”
650 30’ 00” * 3 = 1960 30’ 00”
650 30’ 20” * 4 = 2620 01’ 20”
650 30’ 10” * 3 = 1960 30’ 30”
Sum = 9820 31’ 40”
NORMAL EQUATIONS
A normal equation is the one which is formed by multiplying each equation by the coefficient of
the unknown whose normal equation is to be found out by adding the equation thus formed. As
the number of normal equations is the same as the number of unknowns, the most probable
values of the unknowns can be found from the equations.
If the observations are of equal weight, we derive the following rule for forming the normal
equations:
Rule 1. To form a normal equation for each of the unknowns quantities, multiply each
observation equation by the algebraic co-efficient of that unknown quantity in that
equation, and add the results.
If the observation equations are of different weights, we derive the following rule for forming the
normal equations:
Rule 2. To form the normal equation for each of the unknown quantities, multiply each
observation equation by the product of the algebraic coefficient of that unknown quantity
in that equation and the weight of that observation and add the results.
Example 5: (a) From the normal equations for x, y, z in the following equations of equal weight:
3x + 3y + z - 4 = 0 ..(1)
x + 2y + 2z - 6 = 0 ..(2)
5x + y + 4z - 21 = 0 ..(3)
(b) If the weights of the above equations are 2, 3, and 1 respectively, from the normal equations
for x, y, z.
Solutions:
(a) The normal equations of unknown quantity is formed by multiplying each equations by
the algebraic coefficient of that unknown quantity in that equation and adding the result.
Thus, in equations (1), (2), (3) the coefficients of x are 3, 1, and 5 respectively
Hence,
9x + 9y+ 3z - 12 = 0
x + 2y+ 2z - 6 = 0
25x + 5y+ 20z - 105 = 0
So Normal equation for x is 35x + 16y + 25z – 123 = 0
(b) The normal equations of unknown quantity is formed by multiplying each equations by
the algebraic coefficient of that quantity in that equation and weight of that equation, and
adding the result.
Thus, in equations (1), (2) and (3) the product of coefficients of x and weight of
respective equations are (3 X2), (1 x 3) and (5 x 1) respectively. Hence
Similarly, the coefficients for y and weight of each equation, in the original equations are (3
X2), (2 x 3) and (1 x 1) respectively. Hence
18x + 18y + 6z – 24 = 0 (from 1)
6x + 12y + 12z – 36 = 0 (from 2)
5x + y + 4z – 21 = 0 (from 3)
So Normal equation for y is 29x + 31y + 22z – 81 = 0
Similarly, the coefficients for y and weight of each equation, in the original equations are (1
X2), (2 x 3) and (4 x 1) respectively. Hence
6x + 6y + 2z – 8 = 0 (from 1)
6x + 12y + 12z – 36 = 0 (from 2)
20x + 4y + 16z – 84 = 0 (from 3)
So Normal equation for y is 32x + 22y + 30z – 128 = 0
The most probable value of a quantity is the one which has more chances of being true than has
any other. It is deduced from several measurements on which it is based. In practice, the
following cases may arise of which the most probable value may be required to be determined:
1. Direct observations of equal weights.
The most probable value of the directly observed quantity of equal weights is equal to the
arithmetic mean of the observed values.
Thus, if V1, V2, V3,…., Vn is the observed value of a quantity of equal weight, and M is the
arithmetic mean, then
M= = most probable value
The most probable value of an observed quantity of unequal weights is equal to the weighted
arithmetic mean of the observed quantities.
Thus, if V1, V2, V3,…., Vn are the observed quantities with weights w1, w2, w3,…., wn and N is
the most probable value of the quantity, we have
N=
When the unknowns are independent of each other, their most probable values can be found by
forming the normal equations for each of the unknowns quantities, and treating them as
simultaneous equations to get the values of the unknowns.
When the observation equations are accompanied by one or more condition equations, the latter
may be reduced to an observation equation which will eliminate one of the unknowns. The
normal equation can then be formed for the remaining unknowns. There is also another method,
known as the method of correlates by which the observation equations are eliminated. However,
the former method (i.e., eliminating the condition equation) is suitable for simple cases while the
latter method is used for more complicated problems.
Example 6: Find the most probable value of the angle A from the following observation
equation:
A = 30˚28'40"
3A = 91˚25'55"
4A = 121˚54'30"
Solution:
There is only one unknown, and all the observations are of equal weight. The coefficient of A in
three equations are 1, 3 and 4. Hence multiply these equations by 1, 3 and 4 respectively and add
the resulting equations to get normal equation for A.
Thus,
A = 30˚28'40"
9A = 274˚17'45"
16A = 487˚38'00"
26A = 792˚24'25" (Normal equation in A)
A = 30˚ 28' 37.9"
Alternative solution
Weight = 1/ (1/3)2 = 9
From third equation A = (121˚54'30")/4 = 30˚28'37.5"
Weight = 1/ (1/4)2
Sum of weights = 1+9+16 =26
Weighted mean (A) = 30˚28' + [(40x1) +(38.33x9)+(37.5x16)]/26"
= 30˚ 28' 37.9"
Example 7: Find the most probable value of the angle A from the following observation
equations:
A = 30˚28'40" weight 2.
3A = 91˚25'55" weight 3.
Solution:
There is only one unknown, and however the observations are of unequal weight. The normal
equation can be formed by multiplying each of the two observation equation by the
corresponding weight and coefficient of A, and adding them.
2x1xA = 2A = 60˚57'20"
3x3x3A = 27A = 822˚53'15"
29A = 883˚50'35" (Normal equation in A)
A = 30˚28'38.5"
Example 8: Find the most probable value of the angles A and B from the following observations
at a station O:
A = 49˚48'36.6" weight 2.
B = 54˚37'48.3" weight 3.
A+B = 104˚26'28.5" weight 4.
Solution:
There are two unknown A and B and both are independent of each other, and there will be two
normal equations.
To find normal equation for A,
2A = 99˚37'13.2"
4A+4B = 417˚45'54"
6A+4B = 517˚23'7.2" (Normal equation for A)
To find normal equation for B,
3B = 163˚53'24.9"
4A+4B = 417˚45'54"
4A+7B = 581˚39'18.9" (Normal equation for B)
Hence normal equations are,
6A + 4B = 517˚23'7.2"
4A + 7B = 581˚39'18.9"
To solve A and B, multiply the equations by 2 and 3 respectively. Thus,
12A + 8B = 1034˚46'14.4" …(1)
12A + 21B = 1744˚57'56.7" …(2)
Example 9: The following are the mean value observed in the measurement of three angles α, β
and γ at one station:
α = 76˚42'46.2" weight 4.
α + β = 134˚36'32.6" weight 3.
β + γ = 185˚35'24.8" weight 2.
α + β + γ = 262˚18'10.4" weight 1.
Calculate the most probable value of each angle.
Solution:
To form the normal equation for the unknown, multiply each equation by the coefficient of that
unknown and also by the weight of the equation, and take the sum of resulting equations.
Thus, forming normal equations for α, we have
4α = 306˚51'4.8"
3α + 3β = 403˚49'37.8"
α + β + γ = 262˚18'10.4"
8α + 4β + γ = 972˚58'53" (normal equation for α)
Forming normal equations for β, we have
3α + 3β = 403˚49'37.8"
2β + 2γ = 371˚10'49.6"
α + β + γ = 262˚18'10.4"
4α + 6β + 3γ = 1037˚18'10.4" (normal equation for β)
Forming normal equations for γ, we have
2β + 2γ = 371˚10'49.6"
α + β + γ = 262˚18'10.4"
α + 3β + 3γ = 633˚29'00" (normal equation for γ)
Hence the three normal equations are:
8α + 4β + γ = 972˚58'53"
4α + 6β + 3γ = 1037˚18'10.4"
α + 3β + 3γ = 633˚29'00"
Solving the above three equations simultaneously for α, β and γ we get
α = 76˚ 42' 46.17"
β = 57˚ 53' 46.13"
γ = 127˚ 41' 38.26"
Example 10: The following are the observed values of A, B and C at a station. The angle being
subjected to the condition that A + B = C
A = 30 12’ 28’’.2
B = 35 48’ 12’’.6
C = 66 0’ 44’’.4
Find the most probable values of A, B and C.
Solution:
Example 11: Find the most probable values of angles A, B and C of triangle ABC from the
following observation equations
A = 68 12’ 36’’
B = 53 46’ 12’’
C = 58 01’ 16’’
Solution:
C = 1800 – (A + B) = 1800 – (770 14’ 23” + 490 40’ 39”) = 530 4’ 58”