Search

Autonomous vehicles are rapidly advancing and have the potential to revolutionize transportation in the future. This paper primarily focuses on vehicle motion trajectory planning algorithms, examining the methods for estimating collision risks based on sensed environmental information and approaches for achieving user-aligned trajectory planning results. It investigates the different categories of planning algorithms within the scope of local trajectory planning applications for autonomous driving, discussing and differentiating their properties in detail through a review of the recent studies. The risk estimation methods are classified and introduced based on their descriptions of the sensed collision risks in traffic environments and their integration with trajectory planning algorithms. Additionally, various user experience-oriented methods, which utilize human data to enhance the trajectory planning performance and generate human-like trajectories, are explored. The paper provides comparative analyses of these algorithms and methods from different perspectives, revealing the interconnections between these topics. The current challenges and future prospects of the trajectory planning tasks in autonomous vehicles are also discussed. [ABSTRACT FROM AUTHOR]

Object detection is one of the core technologies for autonomous driving. Current road object detection mainly relies on visible light, which is prone to missed detections and false alarms in rainy, night-time, and foggy scenes. Multispectral object detection based on the fusion of RGB and infrared images can effectively address the challenges of complex and changing road scenes, improving the detection performance of current algorithms in complex scenarios. However, previous multispectral detection algorithms suffer from issues such as poor fusion of dual-mode information, poor detection performance for multi-scale objects, and inadequate utilization of semantic information. To address these challenges and enhance the detection performance in complex road scenes, this paper proposes a novel multispectral object detection algorithm called MRD-YOLO. In MRD-YOLO, we utilize interaction-based feature extraction to effectively fuse information and introduce the BIC-Fusion module with attention guidance to fuse different modal information. We also incorporate the SAConv module to improve the model's detection performance for multi-scale objects and utilize the AIFI structure to enhance the utilization of semantic information. Finally, we conduct experiments on two major public datasets, FLIR_Aligned and M3FD. The experimental results demonstrate that compared to other algorithms, the proposed algorithm achieves superior detection performance in complex road scenes. [ABSTRACT FROM AUTHOR]

This paper presents the development of path-tracking control strategies for an over-actuated autonomous electric vehicle. The vehicle platform is equipped with four-wheel steering (4WS) as well as torque vectoring (TV) capabilities, which enable the control of vehicle dynamics to be enhanced. A nonlinear model predictive controller is proposed taking into account the nonlinearities in vehicle dynamics at the limits of handling as well as the crucial actuator constraints. Controllers with different actuation formulations are presented and compared to study the path-tracking performance of the vehicle with different levels of actuation. The controllers are implemented in a high-fidelity simulation environment considering scenarios of vehicle handling limits. According to the simulation results, the vehicle achieves the best overall path-tracking performance with combined 4WS and TV, which illustrates that the over-actuation topology can enhance the path-tracking performance during conditions under the limits of handling. In addition, the performance of the over-actuation controller is further assessed with different sampling times as well as prediction horizons in order to investigate the effect of such parameters on the control performance, and its capability for real-time execution. In the end, the over-actuation control strategy is implemented on a target machine for real-time validation. The control formulation proposed in this paper is proven to be compatible with different levels of actuation, and it is also demonstrated in this work that it is possible to include the particular over-actuation formulation and specific nonlinear vehicle dynamics in real-time operation, with the sampling time and prediction time providing a compromise between path-tracking performance and computational time. [ABSTRACT FROM AUTHOR]

This paper considers the interactive effects between the ego vehicle and other vehicles in a dynamic driving environment and proposes an autonomous vehicle lane-changing behavior decision-making and trajectory planning method based on graph convolutional networks (GCNs) and multi-segment polynomial curve optimization. Firstly, hierarchical modeling is applied to the dynamic driving environment, aggregating the dynamic interaction information of driving scenes in the form of graph-structured data. Graph convolutional neural networks are employed to process interaction information and generate ego vehicle's driving behavior decision commands. Subsequently, collision-free drivable areas are constructed based on the dynamic driving scene information. An optimization-based multi-segment polynomial curve trajectory planning method is employed to solve the optimization model, obtaining collision-free motion trajectories satisfying dynamic constraints and efficiently completing the lane-changing behavior of the vehicle. Finally, simulation and on-road vehicle experiments are conducted for the proposed method. The experimental results demonstrate that the proposed method outperforms traditional decision-making and planning methods, exhibiting good robustness, real-time performance, and strong scenario generalization capabilities. [ABSTRACT FROM AUTHOR]

For the RRT* algorithm, there are problems such as greater randomness, longer time consumption, more redundant nodes, and inability to perform local obstacle avoidance when encountering unknown obstacles in the path planning process of autonomous vehicles. And the artificial potential field method (APF) applied to autonomous vehicles is prone to problems such as local optimality, unreachable targets, and inapplicability to global scenarios. A fusion algorithm combining the improved RRT* algorithm and the improved artificial potential field method is proposed. First of all, for the RRT* algorithm, the concept of the artificial potential field and probability sampling optimization strategy are introduced, and the adaptive step size is designed according to the road curvature. The path post-processing of the planned global path is carried out to reduce the redundant nodes of the generated path, enhance the purpose of sampling, solve the problem where oscillation may occur when expanding near the target point, reduce the randomness of RRT* node sampling, and improve the efficiency of path generation. Secondly, for the artificial potential field method, by designing obstacle avoidance constraints, adding a road boundary repulsion potential field, and optimizing the repulsion function and safety ellipse, the problem of unreachable targets can be solved, unnecessary steering in the path can be reduced, and the safety of the planned path can be improved. In the face of U-shaped obstacles, virtual gravity points are generated to solve the local minimum problem and improve the passing performance of the obstacles. Finally, the fusion algorithm, which combines the improved RRT* algorithm and the improved artificial potential field method, is designed. The former first plans the global path, extracts the path node as the temporary target point of the latter, guides the vehicle to drive, and avoids local obstacles through the improved artificial potential field method when encountered with unknown obstacles, and then smooths the path planned by the fusion algorithm, making the path satisfy the vehicle kinematic constraints. The simulation results in the different road scenes show that the method proposed in this paper can quickly plan a smooth path that is more stable, more accurate, and suitable for vehicle driving. [ABSTRACT FROM AUTHOR]

The paper demonstrates the validity of sensors and the model in the algorithm for a lane change controller. The paper presents the systematic derivation of the chosen model from the ground up and the important role played by the sensors used in this system. The whole concept of the system on which the tests were carried out is presented step by step. Simulations were realised in the Matlab and Simulink environments. Preliminary tests were performed to confirm the need for the controller in a closed-loop system. On the other hand, sensitivity (the influence of noise and offset) studies showed the advantages and disadvantages of the developed algorithm. This allowed us to create a research path for future work with the aim of improving the operation of the proposed system. [ABSTRACT FROM AUTHOR]

Complete autonomous systems such as self-driving cars to ensure the high reliability and safety of humans need the most efficient combination of four-dimensional (4D) detection, exact localization, and artificial intelligent (AI) networking to establish a fully automated smart transportation system. At present, multiple integrated sensors such as light detection and ranging (LiDAR), radio detection and ranging (RADAR), and car cameras are frequently used for object detection and localization in the conventional autonomous transportation system. Moreover, the global positioning system (GPS) is used for the positioning of autonomous vehicles (AV). These individual systems' detection, localization, and positioning efficiency are insufficient for AV systems. In addition, they do not have any reliable networking system for self-driving cars carrying us and goods on the road. Although the sensor fusion technology of car sensors came up with good efficiency for detection and location, the proposed convolutional neural networking approach will assist to achieve a higher accuracy of 4D detection, precise localization, and real-time positioning. Moreover, this work will establish a strong AI network for AV far monitoring and data transmission systems. The proposed networking system efficiency remains the same on under-sky highways as well in various tunnel roads where GPS does not work properly. For the first time, modified traffic surveillance cameras have been exploited in this conceptual paper as an external image source for AV and anchor sensing nodes to complete AI networking transportation systems. This work approaches a model that solves AVs' fundamental detection, localization, positioning, and networking challenges with advanced image processing, sensor fusion, feathers matching, and AI networking technology. This paper also provides an experienced AI driver concept for a smart transportation system with deep learning technology. [ABSTRACT FROM AUTHOR]

In this paper, the low-level velocity controller of an autonomous vehicle is studied. The performance of the traditional controller used in this kind of system, a PID, is analyzed. This kind of controller cannot follow ramp references without error, so when the reference implies a change in the speed, the vehicle cannot follow the proposed reference, and there is a significant difference between the actual and desired vehicle behaviors. A fractional controller is proposed which changes the ordinary dynamics allowing faster responses for small times, at the cost of slower responses for large times. The idea is to take advantage of this fact to follow fast setpoint changes with a smaller error than that obtained with a classic non-fractional PI controller. Using this controller, the vehicle can follow variable speed references with zero stationary error, significantly reducing the difference between reference and actual vehicle behavior. The paper presents the fractional controller, studies its stability in function of the fractional parameters, designs the controller, and tests its stability. The designed controller is tested on a real prototype, and its behavior is compared to a standard PID controller. The designed fractional PID controller overcomes the results of the standard PID controller. [ABSTRACT FROM AUTHOR]

Intersections are at the core of congestion in urban areas. After the end of the Second World War, the problem of intersection management has benefited from a growing body of advances to address the optimization of the traffic lights' phase splits, timing, and offset. These contributions have significantly improved traffic safety and efficiency in urban areas. However, with the growth of transportation demand and motorization, traffic lights show their limits. At the end of the 1990s, the perspective of autonomous and connected driving systems motivated researchers to introduce a paradigm shift for controlling intersections. This new paradigm is well known today as autonomous intersection management (AIM). It harnesses the self-organization ability of future vehicles to provide more accurate control approaches that use the smallest available time window to reach unprecedented traffic performances. This is achieved by optimizing two main points of the interaction of connected and autonomous vehicles at intersections: the motion control of vehicles and the schedule of their accesses. Considering the great potential of AIM and the complexity of the problem, the proposed approaches are very different, starting from various assumptions. With the increasing popularity of AIM, this paper provides readers with a comprehensive vision of noticeable advances toward enhancing traffic efficiency. It shows that it is possible to tailor vehicles' speed and schedule according to the traffic demand by using distributed particle swarm optimization. Moreover, it brings the most relevant contributions in the light of traffic engineering, where flow–speed diagrams are used to measure the impact of the proposed optimizations. Finally, this paper presents the current challenging issues to be addressed. [ABSTRACT FROM AUTHOR]

Autonomous driving technology plays an essential role in reducing road traffic accidents and ensuring more convenience while driving, so it has been widely studied in industrial and academic communities. The lane-changing decision-making process is challenging but critical for ensuring autonomous vehicles' (AVs) safe and smooth maneuvering. This paper presents a closed-loop lane-changing behavioral decision-making framework suitable for AVs in fully autonomous driving environments to achieve both safety and high efficiency. The framework is based on a complete information non-cooperative game theory. Moreover, we attempt to introduce human driver-specific driving styles (reflected by aggressiveness types) and micro-interaction behaviors for both sides of the game in this model, enabling users to understand, adapt, and utilize intelligent lane-changing techniques. Additionally, a model predictive control controller based on the host-vehicle (HV) driving risk field (DRF) is proposed. The controller's optimizer is used to find the optimal path with the lowest driving risk by using its optimizer and simultaneously adjusting its control variables to track the path. The method can synchronize path planning and motion control and provide real-time vehicle state feedback to the decision-making module. Simulations in several typical traffic scenarios demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

With the advent of autonomous vehicles, sensors and algorithm testing have become crucial parts of the autonomous vehicle development cycle. Having access to real-world sensors and vehicles is a dream for researchers and small-scale original equipment manufacturers (OEMs) due to the software and hardware development life-cycle duration and high costs. Therefore, simulator-based virtual testing has gained traction over the years as the preferred testing method due to its low cost, efficiency, and effectiveness in executing a wide range of testing scenarios. Companies like ANSYS and NVIDIA have come up with robust simulators, and open-source simulators such as CARLA have also populated the market. However, there is a lack of lightweight and simple simulators catering to specific test cases. In this paper, we introduce the SLAV-Sim, a lightweight simulator that specifically trains the behaviour of a self-learning autonomous vehicle. This simulator has been created using the Unity engine and provides an end-to-end virtual testing framework for different reinforcement learning (RL) algorithms in a variety of scenarios using camera sensors and raycasts. [ABSTRACT FROM AUTHOR]

Ensuring that intelligent vehicles do not cause fatal collisions remains a persistent challenge due to pedestrians' unpredictable movements and behavior. The potential for risky situations or collisions arising from even minor misunderstandings in vehicle–pedestrian interactions is a cause for great concern. Considerable research has been dedicated to the advancement of predictive models for pedestrian behavior through trajectory prediction, as well as the exploration of the intricate dynamics of vehicle–pedestrian interactions. However, it is important to note that these studies have certain limitations. In this paper, we propose a novel graph-based trajectory prediction model for vehicle–pedestrian interactions called Holistic Spatio-Temporal Graph Attention (HSTGA) to address these limitations. HSTGA first extracts vehicle–pedestrian interaction spatial features using a multi-layer perceptron (MLP) sub-network and max pooling. Then, the vehicle–pedestrian interaction features are aggregated with the spatial features of pedestrians and vehicles to be fed into the LSTM. The LSTM is modified to learn the vehicle–pedestrian interactions adaptively. Moreover, HSTGA models temporal interactions using an additional LSTM. Then, it models the spatial interactions among pedestrians and between pedestrians and vehicles using graph attention networks (GATs) to combine the hidden states of the LSTMs. We evaluate the performance of HSTGA on three different scenario datasets, including complex unsignalized roundabouts with no crosswalks and unsignalized intersections. The results show that HSTGA outperforms several state-of-the-art methods in predicting linear, curvilinear, and piece-wise linear trajectories of vehicles and pedestrians. Our approach provides a more comprehensive understanding of social interactions, enabling more accurate trajectory prediction for safe vehicle navigation. [ABSTRACT FROM AUTHOR]

Methods based on 64-beam LiDAR can provide very precise 3D object detection. However, highly accurate LiDAR sensors are extremely costly: a 64-beam model can cost approximately USD 75,000. We previously proposed SLS–Fusion (sparse LiDAR and stereo fusion) to fuse low-cost four-beam LiDAR with stereo cameras that outperform most advanced stereo–LiDAR fusion methods. In this paper, and according to the number of LiDAR beams used, we analyzed how the stereo and LiDAR sensors contributed to the performance of the SLS–Fusion model for 3D object detection. Data coming from the stereo camera play a significant role in the fusion model. However, it is necessary to quantify this contribution and identify the variations in such a contribution with respect to the number of LiDAR beams used inside the model. Thus, to evaluate the roles of the parts of the SLS–Fusion network that represent LiDAR and stereo camera architectures, we propose dividing the model into two independent decoder networks. The results of this study show that—starting from four beams—increasing the number of LiDAR beams has no significant impact on the SLS–Fusion performance. The presented results can guide the design decisions by practitioners. [ABSTRACT FROM AUTHOR]

It is important to estimate the exact depth from 2D images, and many studies have been conducted for a long period of time to solve depth estimation problems. Recently, as research on estimating depth from monocular camera images based on deep learning is progressing, research for estimating accurate depths using various techniques is being conducted. However, depth estimation from 2D images has been a problem in predicting the boundary between objects. In this paper, we aim to predict sophisticated depths by emphasizing the precise boundaries between objects. We propose a depth estimation network with encoder–decoder structures using the Laplacian pyramid and local planar guidance method. In the process of upsampling the learned features using the encoder, the purpose of this step is to obtain a clearer depth map by guiding a more sophisticated boundary of an object using the Laplacian pyramid and local planar guidance techniques. We train and test our models with KITTI and NYU Depth V2 datasets. The proposed network constructs a DNN using only convolution and uses the ConvNext networks as a backbone. A trained model shows the performance of the absolute relative error (Abs_rel) 0.054 and root mean square error (RMSE) 2.252 based on the KITTI dataset and absolute relative error (Abs_rel) 0.102 and root mean square error 0.355 based on the NYU Depth V2 dataset. On the state-of-the-art monocular depth estimation, our network performance shows the fifth-best performance based on the KITTI Eigen split and the eighth-best performance based on the NYU Depth V2. [ABSTRACT FROM AUTHOR]

The automobile industry has developed dramatically in recent years, the supply of vehicles has also increased, and thus it has become deeply established in everyday life. Recently, as the supply of vehicles with autonomous driving functions increases, the safety of vehicles is also an emerging issue. Various car-following models for the safe driving of vehicles have long been studied by various people, and recently, a Responsibility-Sensitive Safety (RSS) model has been proposed by Mobileye. However, in existing car-following models or the RSS model, the safe distance between vehicles is presented using only vehicle speed and acceleration information, so there is a limitation in that it cannot respond to changes in road conditions due to the weather. In this paper, in order to ensure safety when the RSS model is applied to a variable focus function camera, an improved RSS model is presented in consideration of the changes in road conditions due to changes in weather, and a safety distance is derived based on the proposed model. [ABSTRACT FROM AUTHOR]

This paper presents a sliding mode control (SMC)-based path-tracking algorithm for autonomous vehicles by considering model-free adaptive feedback actions. In autonomous vehicles, safe path tracking requires adaptive and robust control algorithms because driving environment and vehicle conditions vary in real time. In this study, the SMC was adopted as a robust control method to adjust the switching gain, taking into account the sliding surface and unknown uncertainty to make the control error zero. The sliding surface can be designed mathematically, but it is difficult to express the unknown uncertainty mathematically. Information of priori bounded uncertainties is needed to obtain closed-loop stability of the control system, and the unknown uncertainty can vary with changes in internal and external factors. In the literature, ongoing efforts have been made to overcome the limitation of losing control stability due to unknown uncertainty. This study proposes an integrated method of adaptive feedback control (AFC) and SMC that can adjust a bounded uncertainty. Some illustrative and representative examples, such as autonomous driving scenarios, are also provided to show the main properties of the designed integrated controller. The examples show superior control performance, and it is expected that the integrated controller could be widely used for the path-tracking algorithms of autonomous vehicles. [ABSTRACT FROM AUTHOR]

This paper presents an interactive lane keeping model for an advanced driver assistant system and autonomous vehicle. The proposed model considers not only the lane markers but also the interaction with surrounding vehicles in determining steering inputs. The proposed algorithm is designed based on the Recurrent Neural Network (RNN) with long short-term memory cells, which are configured by the collected driving data. A data collection vehicle is equipped with a front camera, LiDAR, and DGPS. The input features of the RNN consist of lane information, surrounding targets, and ego vehicle states. The output feature is the steering wheel angle to keep the lane. The proposed algorithm is evaluated through similarity analysis and a case study with driving data. The proposed algorithm shows accurate results compared to the conventional algorithm, which only considers the lane markers. In addition, the proposed algorithm effectively responds to the surrounding targets by considering the interaction with the ego vehicle. [ABSTRACT FROM AUTHOR]

This paper presents a novel vehicular environment identification approach based on deep learning. It consists of exploiting the vehicular wireless channel characteristics in the form of Channel State Information (CSI) in the receiver side of a connected vehicle in order to identify the environment type in which the vehicle is driving, without any need to implement specific sensors such as cameras or radars. We consider environment identification as a classification problem, and propose a new convolutional neural network (CNN) architecture to deal with it. The estimated CSI is used as the input feature to train the model. To perform the identification process, the model is targeted for implementation in an autonomous vehicle connected to a vehicular network (VN). The proposed model is extensively evaluated, showing that it can reliably recognize the surrounding environment with high accuracy ( 96.48 % ). Our model is compared to related approaches and state-of-the-art classification architectures. The experiments show that our proposed model yields favorable performance compared to all other considered methods. [ABSTRACT FROM AUTHOR]

In this paper, we propose a new free space detection algorithm for autonomous vehicle driving. Previous free space detection algorithms often use only the location information of every frame, without information on the speed of the obstacle. In this case, there is a possibility of creating an inefficient path because the behavior of the obstacle cannot be predicted. In order to compensate for the shortcomings of the previous algorithm, the proposed algorithm uses the speed information of the obstacle. Through object tracking, the dynamic behavior of obstacles around the vehicle is identified and predicted, and free space is detected based on this. In the free space, it is possible to classify an area in which driving is possible and an area in which it is not possible, and a route is created according to the classification result. By comparing and evaluating the path generated by the previous algorithm and the path generated by the proposed algorithm, it is confirmed that the proposed algorithm is more efficient in generating the vehicle driving path. [ABSTRACT FROM AUTHOR]

The fatigue life of orthotropic steel decks (OSDs) is significantly affected by vehicle loads, and the local stress response of OSDs is sensitive to the transverse position of vehicle loads. However, the presence of autonomous vehicles is likely to change the transverse distribution of vehicles within the lane, thereby affecting vehicle-induced fatigue damage to OSDs. Therefore, it is necessary to evaluate the potential effect of autonomous vehicles on the fatigue life of OSDs so that appropriate strategies can be implemented to control the transverse positions of autonomous vehicles passing the bridge deck. To this end, fatigue damages of several typical fatigue details in a conventional OSD (COSD) and a lightweight composite OSD (LWCD) induced by vehicle loads were calculated based on finite element analysis, and their fatigue lives were evaluated based on Miner's Rule, in which different transverse distribution patterns of autonomous vehicles and their proportions in the mixed traffic flow were considered. The results indicate that fatigue lives of both the COSD and the LWCD can be negatively affected by autonomous vehicles traveling across the bridge without any constraints on the transverse distribution, especially when their proportion in the mixed traffic flow exceeds 30%. Compared to the scenario without autonomous vehicles, the fatigue damage of most fatigue details in OSDs may increase by 51% to 210% in the most unfavorable case due to the presence of autonomous vehicles. Nevertheless, it is feasible to extend the fatigue life of OSDs by optimizing the transverse distribution of autonomous vehicles. Specifically, the fatigue life of most fatigue details in the COSD could be extended by more than 86% in the most favorable case when a bimodal Gaussian distribution is adopted as the transverse distribution pattern of autonomous vehicles. Moreover, both the negative and positive effects of autonomous vehicles on the fatigue life of the COSD are more significant than those of the LWCD in most cases. The results can provide references for the maintenance of OSDs under the action of autonomous vehicles. [ABSTRACT FROM AUTHOR]

With the continuous development of new sensor features and tracking algorithms for object tracking, researchers have opportunities to experiment using different combinations. However, there is no standard or agreed method for selecting an appropriate architecture for autonomous vehicle (AV) crash reconstruction using multi-sensor-based sensor fusion. This study proposes a novel simulation method for tracking performance evaluation (SMTPE) to solve this problem. The SMTPE helps select the best tracking architecture for AV crash reconstruction. This study reveals that a radar-camera-based centralized tracking architecture of multi-sensor fusion performed the best among three different architectures tested with varying sensor setups, sampling rates, and vehicle crash scenarios. We provide a brief guideline for the best practices in selecting appropriate sensor fusion and tracking architecture arrangements, which can be helpful for future vehicle crash reconstruction and other AV improvement research. [ABSTRACT FROM AUTHOR]

Road and sidewalk detection in urban scenarios is a challenging task because of the road imperfections and high sensor data bandwidth. Traditional free space and ground filter algorithms are not sensitive enough for small height differences. Camera-based or sensor-fusion solutions are widely used to classify drivable road from sidewalk or pavement. A LIDAR sensor contains all the necessary information from which the feature extraction can be done. Therefore, this paper focuses on LIDAR-based feature extraction. For road and sidewalk detection, the current paper presents a real-time (20 Hz+) solution. This solution can also be used for local path planning. Sidewalk edge detection is the combination of three algorithms working parallelly. To validate the result, the de facto standard benchmark dataset, KITTI, was used alongside our measurements. The data and the source code to reproduce the results are shared publicly on our GitHub repository. [ABSTRACT FROM AUTHOR]

Autonomous Vehicles (AVs) have the potential to solve many traffic problems, such as accidents, congestion and pollution. However, there are still challenges to overcome, for instance, AVs need to accurately perceive their environment to safely navigate in busy urban scenarios. The aim of this paper is to review recent articles on computer vision techniques that can be used to build an AV perception system. AV perception systems need to accurately detect non-static objects and predict their behaviour, as well as to detect static objects and recognise the information they are providing. This paper, in particular, focuses on the computer vision techniques used to detect pedestrians and vehicles. There have been many papers and reviews on pedestrians and vehicles detection so far. However, most of the past papers only reviewed pedestrian or vehicle detection separately. This review aims to present an overview of the AV systems in general, and then review and investigate several detection computer vision techniques for pedestrians and vehicles. The review concludes that both traditional and Deep Learning (DL) techniques have been used for pedestrian and vehicle detection; however, DL techniques have shown the best results. Although good detection results have been achieved for pedestrians and vehicles, the current algorithms still struggle to detect small, occluded, and truncated objects. In addition, there is limited research on how to improve detection performance in difficult light and weather conditions. Most of the algorithms have been tested on well-recognised datasets such as Caltech and KITTI; however, these datasets have their own limitations. Therefore, this paper recommends that future works should be implemented on more new challenging datasets, such as PIE and BDD100K. [ABSTRACT FROM AUTHOR]

With the emergence of autonomous functions in road vehicles, there has been increased use of Advanced Driver Assistance Systems comprising various sensors to perform automated tasks. Light Detection and Ranging (LiDAR) is one of the most important types of optical sensor, detecting the positions of obstacles by representing them as clusters of points in three-dimensional space. LiDAR performance degrades significantly when a vehicle is driving in the rain as raindrops adhere to the outer surface of the sensor assembly. Performance degradation behaviors include missing points and reduced reflectivity of the points. It was found that the extent of degradation is highly dependent on the interface material properties. This subsequently affects the shapes of the adherent droplets, causing different perturbations to the optical rays. A fundamental investigation is performed on the protective polycarbonate cover of a LiDAR assembly coated with four classes of material—hydrophilic, almost-hydrophobic, hydrophobic, and superhydrophobic. Water droplets are controllably dispensed onto the cover to quantify the signal alteration due to the different droplets of various sizes and shapes. To further understand the effects of droplet motion on LiDAR signals, sliding droplet conditions are simulated using numerical analysis. The results are validated with physical optical tests, using a 905 nm laser source and receiver to mimic the LiDAR detection mechanism. Comprehensive explanations of LiDAR performance degradation in rain are presented from both material and optical perspectives. These can aid component selection and the development of signal-enhancing strategies for the integration of LiDARs into vehicle designs to minimize the impact of rain. [ABSTRACT FROM AUTHOR]

Detecting road cracks is essential for inspecting and assessing the integrity of concrete pavement structures. Traditional image-based methods often require complex preprocessing to extract crack features, making them challenging when dealing with noisy concrete surfaces in diverse real-world scenarios, such as autonomous vehicle road detection. This study introduces an image-based crack detection approach that combines a Random Forest machine learning classifier with a deep convolutional neural network (CNN) to address these challenges. Three state-of-the-art models, namely MobileNet, InceptionV3, and Xception, were employed and trained using a dataset of 30,000 images to build an effective CNN. A systematic comparison of validation accuracy across various base learning rates identified a base learning rate of 0.001 as optimal, achieving a maximum validation accuracy of 99.97%. This optimal learning rate was then applied in the subsequent testing phase. The robustness and flexibility of the trained models were evaluated using 6,000 test photos, each with a resolution of 224 × 224 pixels, which were not part of the training or validation sets. The outstanding results, boasting a remarkable 99.95% accuracy, 99.95% precision, 99.94% recall, and a matching 99.94% F1 Score, unequivocally affirm the efficacy of the proposed technique in precisely identifying road fractures in photographs taken on real concrete surfaces. [ABSTRACT FROM AUTHOR]

Pedestrian detection is a critical task for safety-critical systems, but detecting pedestrians is challenging in low-light and adverse weather conditions. Thermal images can be used to improve robustness by providing complementary information to RGB images. Previous studies have shown that multi-modal feature fusion using convolution operation can be effective, but such methods rely solely on local feature correlations, which can degrade the performance capabilities. To address this issue, we propose an attention-based novel fusion network, referred to as INSANet (INtra-INter Spectral Attention Network), that captures global intra- and inter-information. It consists of intra- and inter-spectral attention blocks that allow the model to learn mutual spectral relationships. Additionally, we identified an imbalance in the multispectral dataset caused by several factors and designed an augmentation strategy that mitigates concentrated distributions and enables the model to learn the diverse locations of pedestrians. Extensive experiments demonstrate the effectiveness of the proposed methods, which achieve state-of-the-art performance on the KAIST dataset and LLVIP dataset. Finally, we conduct a regional performance evaluation to demonstrate the effectiveness of our proposed network in various regions. [ABSTRACT FROM AUTHOR]

Autonomous parking valet systems improve users' comfort, helping with the task of searching for a parking space and parking maneuvering; and due to the simple infrastructure design and low speeds, this maneuver is quite feasible for automated vehicles. Various demonstrations have been performed in both closed parking and in open air parking; scenarios that allow the use of specific technological tools for navigation and searching for a parking space. However, there are still challenges. The purpose of this paper was the integration of perception, positioning, decision-making, and maneuvering algorithms for the control of an autonomous vehicle in a parking lot with the support of a single LiDAR sensor, and with no additional sensors in the infrastructure. Based on a digital map, which was as simplified as possible, the driver can choose the range of parking spaces in which the vehicle must look for a space. From that moment on, the vehicle moves, looking for free places until an available one in the range selected by the driver is found. Then, the vehicle performs the parking maneuver, choosing between two alternatives to optimize the required space. Tests in a real parking lot, with spaces covered with metallic canopies, showed an accurate behavior. [ABSTRACT FROM AUTHOR]

Electric wheelchairs make it easier for disabled and elderly people to live, move, interact, and participate in society. Moving a wheelchair in open spaces is relatively easy, but in closed and small spaces, maneuvering is difficult. Solutions to such problems for people with disabilities are applicable to a relatively small group of recipients and are mostly custom-made solutions, whose considerable cost is a significant barrier to accessibility. New technologies can provide an opportunity to improve the quality of life of people with disabilities in this aspect. Using selected elements of complex automation and control systems, cost-effective solutions can be created that facilitate the functioning of people with disabilities. This paper presents an analysis of hazards and problems when maneuvering a wheelchair in narrow passageways, as well as the authors' solution to this problem, and the concept and assumptions of a mechatronic anti-collision system based on 2D LiDAR laser scanners. This solution is composed of a proprietary 2D rotating scanner mechanism that ensures the acquisition of 3D images of the environment around the wheelchair. Preliminary tests of this solution yielded promising results. Further research will include miniaturization of the device. [ABSTRACT FROM AUTHOR]

This study presents a novel multimodal heterogeneous perception cross-fusion framework for intelligent vehicles that combines data from millimeter-wave radar and camera to enhance target tracking accuracy and handle system uncertainties. The framework employs a multimodal interaction strategy to predict target motion more accurately and an improved joint probability data association method to match measurement data with targets. An adaptive root-mean-square cubature Kalman filter is used to estimate the statistical characteristics of noise under complex traffic scenarios with varying process and measurement noise. Experiments conducted on a real vehicle platform demonstrate that the proposed framework improves reliability and robustness in challenging environments. It overcomes the challenges of insufficient data fusion utilization, frequent leakage, and misjudgment of dangerous obstructions around vehicles, and inaccurate prediction of collision risks. The proposed framework has the potential to advance the state of the art in target tracking and perception for intelligent vehicles. [ABSTRACT FROM AUTHOR]

Driver's hands on/off detection is very important in current autonomous vehicles for safety. Several studies have been conducted to create a precise algorithm. Although many studies have proposed various approaches, they have some limitations, such as robustness and reliability. Therefore, we propose a deep learning model that utilizes in-vehicle data. We also established a data collection system, which collects in-vehicle data that are auto-labeled for efficient and reliable data acquisition. For a robust system, we devised a confidence logic that prevents outliers' sway. To evaluate our model in more detail, we suggested a new metric to explain the events, considering state transitions. In addition, we conducted an extensive experiment on the new drivers to demonstrate our model's generalization ability. We verified that the proposed system achieved a better performance than in previous studies, by resolving their drawbacks. Our model detected hands on/off transitions in 0.37 s on average, with an accuracy of 95.7%. [ABSTRACT FROM AUTHOR]

One of the major challenges of autonomous vehicles (AV) is their interaction with pedestrians. Unofficial interactions such as gestures, eye contact, waving, and flashing lights are very common behavioral patterns for drivers to express their intent to give priority. In our research we composed a virtual reality experiment for a pedestrian crossing in an urban environment in order to test pedestrians' reactions on an LED light display mounted on a virtual AV. Our main research interest was to investigate whether communication patterns influence the decision making of pedestrians when crossing the road. In a VR environment, four scenarios were created with a vehicle approaching a pedestrian crossing with different speeds and displaying a special red/green sign to pedestrians. Here, 51 persons participating in the experiment had to decide when crossing is safe. Results show that the majority of people indicated they would cross in the time windows when it was actually safe to cross. Male subjects made their decision to cross slightly faster but no significant differences were found in the decision making by gender. It was found that age is not an influencing factor, either. Overall, a quick learning process was experienced proving that explicit communication patterns are self-explaining. [ABSTRACT FROM AUTHOR]

Autonomous driving technology, especially motion planning and the trajectory tracking method, is the foundation of an intelligent interconnected vehicle, which needs to be improved urgently. Currently, research on path planning methods has improved, but few of the current studies consider the vehicle's nonlinear characteristics in the reference model, due to the heavy computational effort. At present, most of the algorithms are designed by a linear vehicle model in order to achieve the real-time performance at the cost of lost accuracy. To achieve a better performance, the dynamics and kinematics characteristics of the vehicle must be simulated, and real-time computing ensured at the same time. In this article, a Takagi–Sugeno fuzzy-model-based closed-loop rapidly exploring random tree algorithm with on-line re-planning process is applied to build the motion planner, which effectively improves the vehicle performance of dynamic obstacle avoidance, and plans the local obstacle avoidance path in line with the dynamic characteristics of the vehicle. A nonlinear vehicle model is integrated into the motion planner design directly. For fast local path planning mission, the Takagi–Sugeno fuzzy modelling method is applied to the modeling process in the planner design, so that the vehicle state can be directly utilized into the path planner to create a feasible path in real-time. The performance of the planner was evaluated by numerical simulation. The results demonstrate that the proposed motion planner can effectively generate a reference trajectory that guarantees driving efficiency with a lower re-planning rate. [ABSTRACT FROM AUTHOR]

Big data already covers intelligent vehicles and is driving the autonomous driving industry's transformation. However, the large amounts of driving data generated will result in complex issues and a huge workload for the test and verification processes of an autonomous driving system. Only effective and precise data extraction and recording aimed at the challenges of low efficiency, poor quality, and a long-time limit for traditional data acquisition can substantially reduce the algorithm development cycle. Based on the premise of driver-dominated vehicle movement, the virtual decision-making of autonomous driving systems under the accompanying state was considered as a reference. Based on a dynamic time warping algorithm and forming a data filtering approach under a dynamic time window, an automatic trigger recording control model for human-vehicle difference feature data was suggested. In this method, the data dimension was minimized, and the efficiency of the data mining was improved. The experimental findings showed that the suggested model decreased recorded invalid data by 75.35% on average and saved about 2.65 TB of data storage space per hour. Compared with industrial-grade methods, it saves an average of 307 GB of storage space per hour. [ABSTRACT FROM AUTHOR]

The existing variants of the rapidly exploring random tree (RRT) cannot be effectively applied in local path planning of the autonomous vehicle and solve the coherence problem of paths between the front and back frames. Thus, an improved heuristic Bi-RRT algorithm is proposed, which is suitable for obstacle avoidance of the vehicle in an unknown dynamic environment. The vehicle constraint considering the driver's driving habit and the obstacle-free direct connection mode of two random trees are introduced. Multi-sampling biased towards the target state reduces invalid searches, and parent node selection with the comprehensive measurement index accelerates the algorithm's execution while making the initial path gentle. The adaptive greedy step size, introducing the target direction, expands the node more effectively. Moreover, path reorganization minimizes redundant path points and makes the path's curvature continuous, and path coherence makes paths between the frames connect smoothly. Simulation analysis clarifies the efficient performance of the proposed algorithm, which can generate the smoothest path within the shortest time compared with the other four algorithms. Furthermore, the experiments on dynamic environments further show that the proposed algorithm can generate a differentiable coherence path, ensuring the ride comfort and stability of the vehicle. [ABSTRACT FROM AUTHOR]

Accurate localization for autonomous vehicle operations is essential in dense urban areas. In order to ensure safety, positioning algorithms should implement fault detection and fallback strategies. While many strategies stop the vehicle once a failure is detected, in this work a new framework is proposed that includes an improved reconfiguration module to evaluate the failure scenario and offer alternative positioning strategies, allowing continued driving in degraded mode until a critical failure is detected. Furthermore, as many failures in sensors can be temporary, such as GPS signal interruption, the proposed approach allows the return to a non-fault state while resetting the alternative algorithms used in the temporary failure scenario. The proposed localization framework is validated in a series of experiments carried out in a simulation environment. Results demonstrate proper localization for the driving task even in the presence of sensor failure, only stopping the vehicle when a fully degraded state is achieved. Moreover, reconfiguration strategies have proven to consistently reset the accumulated drift of the alternative positioning algorithms, improving the overall performance and bounding the mean error. [ABSTRACT FROM AUTHOR]

Features such as safety and longer flight times render lighter-than-air robots strong candidates for indoor navigation applications involving people. However, the existing interactive mobility solutions using such robots lack the capability to follow a long-distance user in a relatively larger indoor space. At the same time, the tracking data delivered to these robots are sensitive to uncertainties in indoor environments such as varying intensities of light and electromagnetic field disturbances. Regarding the above shortcomings, we proposed an ultra-wideband (UWB)-based lighter-than-air indoor robot for user-centered interactive applications. We developed the data processing scheme over a robot operating system (ROS) framework to accommodate the robot's integration needs for a user-centered interactive application. In order to explore the user interaction with the robot at a long-distance, the dual interactions (i.e., user footprint following and user intention recognition) were proposed by equipping the user with a hand-held UWB sensor. Finally, experiments were conducted inside a professional arena to validate the robot's pose tracking in which 3D positioning was compared with the 3D laser sensor, and to reveal the applicability of the user-centered autonomous following of the robot according to the dual interactions. [ABSTRACT FROM AUTHOR]

Presently, most passive safety tests are performed with a precisely specified seat position and carefully seated ATD (anthropomorphic test device) dummies. Facing the development of autonomous vehicles, as well as the need for safety verification during crashes with various seat positions such research is even more urgently needed. Apart from the numerical environment, the existing testing equipment is not validated to perform such an investigation. For example, ATDs are not validated for nonstandard seatback positions, and the most accurate method of such research is volunteer tests. The study presented here was performed on a sled test rig utilizing a 50cc Hybrid III dummy according to a full factorial experiment. In addition, input factors were selected in order to verify a safe test condition for surrogate testing. The measured value was head acceleration, which was used for calculation of a head injury criterion. What was found was an optimal seat angle −117°—at which the head injury criteria had the lowest represented value. Moreover, preliminary body dynamics showed a danger of whiplash occurrence for occupants in a fully-reclined seat. [ABSTRACT FROM AUTHOR]

There are numerous global navigation satellite system-denied regions in urban areas, where the localization of autonomous driving remains a challenge. To address this problem, a high-resolution light detection and ranging (LiDAR) sensor was recently developed. Various methods have been proposed to improve the accuracy of localization using precise distance measurements derived from LiDAR sensors. This study proposes an algorithm to accelerate the computational speed of LiDAR localization while maintaining the original accuracy of lightweight map-matching algorithms. To this end, first, a point cloud map was transformed into a normal distribution (ND) map. During this process, vector-based normal distribution transform, suitable for graphics processing unit (GPU) parallel processing, was used. In this study, we introduce an algorithm that enabled GPU parallel processing of an existing ND map-matching process. The performance of the proposed algorithm was verified using an open dataset and simulations. To verify the practical performance of the proposed algorithm, the real-time serial and parallel processing performances of the localization were compared using high-performance and embedded computers, respectively. The distance root-mean-square error and computational time of the proposed algorithm were compared. The algorithm increased the computational speed of the embedded computer almost 100-fold while maintaining high localization precision. [ABSTRACT FROM AUTHOR]

During the automatic driving of a vehicle, the vehicle's positional information is important for vehicle driving control. If fixed-point land markers such as magnetic markers are used, the vehicle's current position error can be calculated only when a marker is detected while driving, and this error can be used to correct the estimation position. Therefore, correction information is used irregularly and intermittently according to the installation intervals of the magnetic markers and the driving speed. If the detected errors are corrected all at once using the position correction method, discontinuity of the position information can occur. This problem causes instability in the vehicle's route guidance control because the position error fluctuates as the vehicle's speed increases. We devised a time-division position correction method that calculates the error using the absolute position of the magnetic marker, which is estimated when the magnetic marker is detected, along with the absolute position information from the magnetic marker database. Instead of correcting the error at once when the position and heading errors are corrected, the correction is performed by dividing the errors multiple times until the next magnetic marker is detected. This prevents sudden discontinuity of the vehicle position information, and the calculated correction amount is used without loss to obtain stable and continuous position information. We conducted driving tests to compare the performances of the proposed algorithm and conventional methods. We compared the continuity of the position information and the mean error and confirmed the superiority of the proposed method in terms of these aspects. [ABSTRACT FROM AUTHOR]

Although numerous road segmentation studies have utilized vision data, obtaining robust classification is still challenging due to vision sensor noise and target object deformation. Long-distance images are still problematic because of blur and low resolution, and these features make distinguishing roads from objects difficult. This study utilizes light detection and ranging (LiDAR), which generates information that camera images lack, such as distance, height, and intensity, as a reliable supplement to address this problem. In contrast to conventional approaches, additional domain transformation to a bird's eye view space is executed to obtain long-range data with resolutions comparable to those of short-range data. This study proposes a convolutional neural network architecture that processes data transformed to a bird's eye view plane. The network's pathways are split into two parts to resolve calibration errors in the transformed image and point cloud. The network, which has modules that operate sequentially at various scaled dilated convolution rates, is designed to quickly and accurately handle a wide range of data. Comprehensive empirical studies using the Karlsruhe Institute of Technology and Toyota Technological Institute's (KITTI's) road detection benchmarks demonstrate that this study's approach takes advantage of camera and LiDAR information, achieving robust road detection with short runtimes. Our result ranks 22nd in the KITTI's leaderboard and shows real-time performance. [ABSTRACT FROM AUTHOR]