Search Results (240)

The paper presents a vision-based obstacle avoidance strategy for lightweight self-driving cars that can be run on a CPU-only device using a single RGB-D camera. The method consists of two steps: visual perception and path planning. The visual perception part uses ORBSLAM3 enhanced with optical flow to estimate the car’s poses and extract rich texture information from the scene. In the path planning phase, the proposed method employs a method combining a control Lyapunov function and control barrier function in the form of a quadratic program (CLF-CBF-QP) together with an obstacle shape reconstruction process (SRP) to plan safe and stable trajectories. To validate the performance and robustness of the proposed method, simulation experiments were conducted with a car in various complex indoor environments using the Gazebo simulation environment. The proposed method can effectively avoid obstacles in the scenes. The proposed algorithm outperforms benchmark algorithms in achieving more stable and shorter trajectories across multiple simulated scenes. Full article

High-precision simultaneous localization and mapping (SLAM) in dynamic real-world environments plays a crucial role in autonomous robot navigation, self-driving cars, and drone control. To address this dynamic localization issue, in this paper, a dynamic odometry method is proposed based on FAST-LIVO, a fast LiDAR (light detection and ranging)–inertial–visual odometry system, integrating neural networks with laser, camera, and inertial measurement unit modalities. The method first constructs visual–inertial and LiDAR–inertial odometry subsystems. Then, a lightweight neural network is used to remove dynamic elements from the visual part, and dynamic clustering is applied to the LiDAR part to eliminate dynamic environments, ensuring the reliability of the remaining environmental data. Validation of the datasets shows that the proposed multi-sensor fusion dynamic odometry can achieve high-precision pose estimation in complex dynamic environments with high continuity, reliability, and dynamic robustness. Full article

Connected cooperative and automated (CAM) vehicles and self-driving cars need to achieve robust and accurate environment understanding. With this aim, they are usually equipped with sensors and adopt multiple sensing strategies, also fused among them to exploit their complementary properties. In recent years, artificial intelligence such as machine learning- and deep learning-based approaches have been applied for object and pedestrian detection and prediction reliability quantification. This paper proposes a procedure based on the YOLOv8 (You Only Look Once) method to discover objects on the roads such as cars, traffic lights, pedestrians and street signs in foggy weather conditions. In particular, YOLOv8 is a recent release of YOLO, a popular neural network model used for object detection and image classification. The obtained model is applied to a dataset including about 4000 foggy road images and the object detection accuracy is improved by changing hyperparameters such as epochs, batch size and augmentation methods. To achieve good accuracy and few errors in detecting objects in the images, the hyperparameters are optimized by four different methods, and different metrics are considered, namely accuracy factor, precision, recall, precision–recall and loss. Full article

Traditional approaches to safety event analysis in autonomous systems have relied on complex machine and deep learning models and extensive datasets for high accuracy and reliability. However, the emerge of multimodal large language models (MLLMs) offers a novel approach by integrating textual, visual, and audio modalities. Our framework leverages the logical and visual reasoning power of MLLMs, directing their output through object-level question–answer (QA) prompts to ensure accurate, reliable, and actionable insights for investigating safety-critical event detection and analysis. By incorporating models like Gemini-Pro-Vision 1.5, we aim to automate safety-critical event detection and analysis along with mitigating common issues such as hallucinations in MLLM outputs. The results demonstrate the framework’s potential in different in-context learning (ICT) settings such as zero-shot and few-shot learning methods. Furthermore, we investigate other settings such as self-ensemble learning and a varying number of frames. The results show that a few-shot learning model consistently outperformed other learning models, achieving the highest overall accuracy of about 79%. The comparative analysis with previous studies on visual reasoning revealed that previous models showed moderate performance in driving safety tasks, while our proposed model significantly outperformed them. To the best of our knowledge, our proposed MLLM model stands out as the first of its kind, capable of handling multiple tasks for each safety-critical event. It can identify risky scenarios, classify diverse scenes, determine car directions, categorize agents, and recommend the appropriate actions, setting a new standard in safety-critical event management. This study shows the significance of MLLMs in advancing the analysis of naturalistic driving videos to improve safety-critical event detection and understanding the interactions in complex environments. Full article

The predictors of urban trip mode choice and one of its important components, public transit ridership, have still not been thoroughly investigated using case studies in Central Europe. Therefore, this study attempts to clarify the correlates of mode choices for commute travel and shopping, and entertainment travel to distant places, as well as the frequencies of public transit use of university students, using a wide range of explanatory variables covering individual, household, and socio-economic attributes as well as their perceptions, mobility, and the nearby built environment. The correlation hypothesis of these factors, especially the role of the street network, was tested by collecting the data from 1288 university students in Krakow and developing Binary Logistic and Ordinal Probit models. The results show that gender, age, car ownership, main daily activity, possession of a driving license, gross monthly income, duration of living in the current home, daily shopping area, sense of belonging to the neighborhood, quality of social/recreational facilities of the neighborhood, and commuting distance can predict commute and non-commute mode choices, while gender, daily activity, financial dependence from the family, entertainment place, quality of social/recreational facilities, residential self-selection, number of commute trips, time living in the current home, and street connectivity around home are significantly correlated with public transit use. Some of these findings are somewhat different from those regarding university students in Western Europe or other high-income countries. These results can be used for policy making to reduce students’ personal and household car use and increase sustainable modal share in Poland and similar neighboring countries. Full article

Artificial intelligence (AI) systems of autonomous systems such as drones, robots and self-driving cars may consume up to 50% of the total power available onboard, thereby limiting the vehicle’s range of functions and considerably reducing the distance the vehicle can travel on a single charge. Next-generation onboard AI systems need an even higher power since they collect and process even larger amounts of data in real time. This problem cannot be solved using traditional computing devices since they become more and more power-consuming. In this review article, we discuss the perspectives on the development of onboard neuromorphic computers that mimic the operation of a biological brain using the nonlinear–dynamical properties of natural physical environments surrounding autonomous vehicles. Previous research also demonstrated that quantum neuromorphic processors (QNPs) can conduct computations with the efficiency of a standard computer while consuming less than 1% of the onboard battery power. Since QNPs are a semi-classical technology, their technical simplicity and low cost compared to quantum computers make them ideally suited for applications in autonomous AI systems. Providing a perspective on the future progress in unconventional physical reservoir computing and surveying the outcomes of more than 200 interdisciplinary research works, this article will be of interest to a broad readership, including both students and experts in the fields of physics, engineering, quantum technologies and computing. Full article

The fifth-generation (5G) cellular infrastructure is expected to bring about the widespread use of connected vehicles. This technological progress marks the beginning of a new era in vehicular networks, which includes a range of different types and services of self-driving cars and the smooth sharing of information between vehicles. Connected vehicles have also been announced as a main use case of the sixth-generation (6G) cellular, with ultimate requirements beyond the 5G (B5G) and 6G eras. These networks require full coverage, extremely high reliability and availability, very low latency, and significant system adaptability. The significant specifications set for vehicular networks pose considerable design and development challenges. The goals of establishing a latency of 1 millisecond, effectively handling large amounts of data traffic, and facilitating high-speed mobility are of utmost importance. To address these difficulties and meet the demands of upcoming networks, e.g., 6G, it is necessary to improve the performance of vehicle networks by incorporating innovative technology into existing network structures. This work presents significant enhancements to vehicular networks to fulfill the demanding specifications by utilizing state-of-the-art technologies, including distributed edge computing, e.g., mobile edge computing (MEC) and fog computing, software-defined networking (SDN), and microservice. The work provides a novel vehicular network structure based on micro-services architecture that meets the requirements of 6G networks. The required offloading scheme is introduced, and a handover algorithm is presented to provide seamless communication over the network. Moreover, a migration scheme for migrating data between edge servers was developed. The work was evaluated in terms of latency, availability, and reliability. The results outperformed existing traditional approaches, demonstrating the potential of our approach to meet the demanding requirements of next-generation vehicular networks. Full article

(1) Background: Few teen driving safety programs focus on increasing parental engagement with high-risk teen drivers, specifically those with a traffic violation. This study explored parents’/guardians’ (‘parents’) experiences with a teen driving safety program, ProjectDRIVE, including facilitators and barriers to program engagement. (2) Methods: We conducted virtual, semi-structured interviews with parents who completed ProjectDRIVE, which included in-vehicle driving feedback technology and individualized virtual training with parents on effective parent–teen communication. (3) Results: Twenty interviews (with 17 females and three males) were transcribed verbatim and independently coded by three coders using systematic, open, and focused coding. Three major themes were identified: factors influencing a parent’s initial decision to participate, factors influencing continued engagement, and perceived benefits of participation. The decision to participate was influenced by these subthemes: parental motivation to help their teen, perceived program usefulness, program endorsement, program incentives, parents’ busy schedules, and lack of access to a car/internet. Subthemes impacting continued engagement included enhanced communication skills, teen willingness to engage, strong parental engagement, and teens’ other priorities. Perceived benefits included greater self-efficacy in communication, improved communication patterns and frequency, and enhanced parent–teen relationships. (4) Conclusions: These findings may set the foundation for developing and implementing future court-ordered parent-based teen safe driving programs for teens with traffic citations. Full article

Rising fossil fuel costs and environmental concerns are driving the search for new energy sources, particularly renewable energy. Among these sources, solar photovoltaic (PV) is the most promising in southern European countries, mainly through the use of decentralised PV systems designed to produce electricity close to the point of demand and primarily to meet local energy needs. In an urban scenario, a decentralised energy system usually operates in parallel with the grid, allowing excess power generated to be injected into the grid. Solar carports and rooftop systems are excellent examples of distributed photovoltaic systems, which are far more sustainable than large centralised systems because they do not compete for land use. Despite their operational advantages, these decentralised photovoltaic production plants, which are in most cases financed by specific energy efficiency programs, present challenges in a regulated market where the injection of energy into the electricity grid is restricted by law and support programs. The aim of this work is to integrate two different photovoltaic systems within an academic campus where the only PV source currently available is a solar car park, a solution designed both to provide shaded space for vehicles and to produce energy to be consumed within the facilities. Due to legal restrictions, surplus electricity cannot be sold to the national grid, and solar batteries to store the generated energy are expensive and have a short lifespan. Therefore, since the campus has two different grid connections and a 102.37 kWp PV system, the newly designed system to be installed on the library roof must be calculated to support the installed electricity system during the most critical working hours, determining the specific angle and orientation of the solar panels. On this basis, the energy management of a school campus is fundamental to creating a collective self-consumption system, the basis of a local energy community that can meet energy, environmental, and social objectives. Full article

Installing multi-camera systems and inertial measurement units (IMUs) in self-driving cars, micro aerial vehicles, and robots is becoming increasingly common. An IMU provides the vertical direction, allowing coordinate frames to be aligned in a common direction. The degrees of freedom (DOFs) of the rotation matrix are reduced from 3 to 1. In this paper, we propose a globally optimal solver to calculate the relative poses and scale of generalized cameras with a known vertical direction. First, the cost function is established to minimize algebraic error in the least-squares sense. Then, the cost function is transformed into two polynomials with only two unknowns. Finally, the eigenvalue method is used to solve the relative rotation angle. The performance of the proposed method is verified on both simulated and KITTI datasets. Experiments show that our method is more accurate than the existing state-of-the-art solver in estimating the relative pose and scale. Compared to the best method among the comparison methods, the method proposed in this paper reduces the rotation matrix error, translation vector error, and scale error by 53%, 67%, and 90%, respectively. Full article

Background: Urban sprawl adversely effects the sustainability of urban environments by promoting private vehicle use, decreasing the viability of active/public transport, and increasing the cost of public service provision. Autonomous vehicles could change the desirability of different residential locations due to resulting changes to urban design and decreased value of travel time. Methods: Adult Australians (n = 1078) completed an online survey that included a description of a future where autonomous vehicles are widely available. The respondents reported anticipated changes in residential location in this autonomous future. Frequency analyses were conducted, and three logistic generalised linear models were run to identify factors associated with staying in the same area or moving to higher- or lower-density locations. Results: Autonomous vehicles are likely to have mixed effects on people’s desired residential locations. Most respondents (84%) elected not to move location, 11% intended to move to lower-density locations, and 6% to higher density locations. Reasons for moving included a desire for more space, the ease of travelling in urban areas, and the reduced value of travel time. Conclusion: The introduction of autonomous vehicles will need to be managed to avoid fostering increased urban sprawl and the associated negative consequences. Strategies that increase the liveability of higher density urban environments are likely to discourage urban sprawl in a future characterised by autonomous transport options. Full article

We investigate personalised feedback mechanisms to help drivers regulate their emotions, aiming to improve road safety. We systematically evaluate driver-preferred feedback modalities and their impact on emotional states. Using unobtrusive vision-based emotion detection and self-labeling, we captured the emotional states and feedback preferences of 21 participants in a simulated driving environment. Results show that in-car feedback systems effectively influence drivers’ emotional states, with participants reporting positive experiences and varying preferences based on their emotions. We also developed a machine learning classification system using facial marker data to demonstrate the feasibility of our approach for classifying emotional states. Our contributions include design guidelines for tailored feedback systems, a systematic analysis of user reactions across three feedback channels with variations, an emotion classification system, and a dataset with labeled face landmark annotations for future research. Full article

Few prior works study self-driving cars by deep learning with IoT collaboration. SDC-Net, which is an end-to-end multitask self-driving car camera cocoon IoT-based system, is one of the research areas that tackles this direction. However, by design, SDC-Net is not able to identify the accident locations; it only classifies whether a scene is a crash scene or not. In this work, we introduce an enhanced design for the SDC-Net system by (1) replacing the classification network with a detection one, (2) adapting our benchmark dataset labels built on the CARLA simulator to include the vehicles’ bounding boxes while keeping the same training, validation, and testing samples, and (3) modifying the shared information via IoT to include the accident location. We keep the same path planning and automatic emergency braking network, the digital automation platform, and the input representations to formulate the comparative study. The SDC-Net++ system is proposed to (1) output the relevant control actions, especially in case of accidents: accelerate, decelerate, maneuver, and brake, and (2) share the most critical information to the connected vehicles via IoT, especially the accident locations. A comparative study is also conducted between SDC-Net and SDC-Net++ with the same input representations: front camera only, panorama and bird’s eye views, and with single-task networks, crash avoidance only, and multitask networks. The multitask network with a BEV input representation outperforms the nearest representation in precision, recall, f1-score, and accuracy by more than 15.134%, 12.046%, 13.593%, and 5%, respectively. The SDC-Net++ multitask network with BEV outperforms SDC-Net multitask with BEV in precision, recall, f1-score, accuracy, and average MSE by more than 2.201%, 2.8%, 2.505%, 2%, and 18.677%, respectively. Full article

Self-driving car research can effectively reduce the occurrence of traffic accidents, but when encountering sudden people or obstacles cutting into the lane, how to reduce the damage hazard to traffic participants and make ethical decisions is the key point that the development of self-driving technology must break through. When faced with sudden traffic participants, self-driving vehicles need to make ethical decisions between ramming into the traffic participants or other obstacles. Therefore, in this paper, we propose a model decision planning method based on multi-objective evaluation function path evaluation of local path planning. This method addresses the ethical model disagreement problem of self-driving vehicles encountering traffic participants and other obstacles. The aim is to ensure the safety of the lives of the traffic participants and achieve the vehicle’s reasonable ethical decision planning. Firstly, when anticipating traffic participants and other obstacles, the vehicle’s planning intention decisions are obtained through fuzzy algorithms. Different sets of curves for various positions are generated based on dynamic programming algorithms. These curves are then fitted using B-spline curves, incorporating obstacle collision costs, and classifying obstacles into different types with varying cost weights. Secondly, factors such as path length and average path curvature are considered for path total cost calculations. Finally, a local path that avoids traffic participants is obtained. This path is then tracked using a pure pursuit algorithm. The proposed algorithm’s effectiveness is verified through simulation experiments and comparative analyses conducted on the MATLAB platform. In conclusion, this research promotes a safer and more sustainable transport system in line with the principles of sustainable development by addressing the challenges associated with safety and ethical decision making in self-driving cars. Full article

Smart cities are now embracing the new frontier of urban living, with advanced technology being used to enhance the quality of life for residents. Many of these cities have developed transportation systems that improve efficiency and sustainability, as well as quality. Integrating cutting-edge transportation technology and data-driven solutions improves safety, reduces environmental impact, optimizes traffic flow during peak hours, and reduces congestion. Intelligent transportation systems consist of many systems, one of which is traffic sign detection. This type of system utilizes many advanced techniques and technologies, such as machine learning and computer vision techniques. A variety of traffic signs, such as yield signs, stop signs, speed limits, and pedestrian crossings, are among those that the traffic sign detection system is trained to recognize and interpret. Ensuring accurate and robust traffic sign recognition is paramount for the safe deployment of self-driving cars in diverse and challenging environments like the Arab world. However, existing methods often face many challenges, such as variability in the appearance of signs, real-time processing, occlusions that can block signs, low-quality images, and others. This paper introduces an advanced Lightweight and Efficient Convolutional Neural Network (LE-CNN) architecture specifically designed for accurate and real-time Arabic traffic sign classification. The proposed LE-CNN architecture leverages the efficacy of depth-wise separable convolutions and channel pruning to achieve significant performance improvements in both speed and accuracy compared to existing models. An extensive evaluation of the LE-CNN on the Arabic traffic sign dataset that was carried out demonstrates an impressive accuracy of 96.5% while maintaining superior performance with a remarkably low inference time of 1.65 s, crucial for real-time applications in self-driving cars. It achieves high accuracy with low false positive and false negative rates, demonstrating its potential for real-world applications like autonomous driving and advanced driver-assistance systems. Full article