Autonomous Vehicle System Research Articles

Transfer Learning-Based Steering Angle Prediction and Control with Fuzzy Signatures-Enhanced Fuzzy Systems for Autonomous Vehicles

This research introduces an innovative approach for End-to-End steering angle prediction and its control in electric power steering (EPS) systems. The methodology integrates transfer learning-based computer vision techniques for prediction and control with fuzzy signatures-enhanced fuzzy systems. Fuzzy signatures are unique multidimensional data structures that represent data symbolically. This enhancement enables the fuzzy systems to effectively manage the inherent imprecision and uncertainty in various driving scenarios. The ultimate goal of this work is to assess the efficiency and performance of this combined approach by highlighting the pivotal role of steering angle prediction and control in the field of autonomous driving systems. Specifically, within EPS systems, the control of the motor directly influences the vehicle’s path and maneuverability. A significant breakthrough of this study is the successful application of transfer learning-based computer vision techniques to extract respective visual data without the need for large datasets. This represents an advancement in reducing the extensive data collection and computational load typically required. The findings of this research reveal the potential of this approach within EPS systems, with an MSE score of 0.0386 against 0.0476, by outperforming the existing NVIDIA model. This result provides a 22.63% better Mean Squared Error (MSE) score than NVIDIA’s model. The proposed model also showed better performance compared with all other three references found in the literature. Furthermore, we identify potential areas for refinement, such as decreasing model loss and simplifying the complex decision model of fuzzy systems, which can represent the symmetry and asymmetry of human decision-making systems. This study, therefore, contributes significantly to the ongoing evolution of autonomous driving systems.

Dataset and Benchmark: Novel Sensors for Autonomous Vehicle Perception

Conventional cameras employed in autonomous vehicle (AV) systems support many perception tasks but are challenged by low-light or high dynamic range scenes, adverse weather, and fast motion. Novel sensors, such as event and thermal cameras, offer capabilities with the potential to address these scenarios, but they remain to be fully exploited. This paper introduces the Novel Sensors for Autonomous Vehicle Perception (NSAVP) dataset to facilitate future research on this topic. The dataset was captured with a platform including stereo event, thermal, monochrome, and RGB cameras as well as a high precision navigation system providing ground truth poses. The data was collected by repeatedly driving two ∼8 km routes and includes varied lighting conditions and opposing viewpoint perspectives. We provide benchmarking experiments on the task of place recognition to demonstrate challenges and opportunities for novel sensors to enhance critical AV perception tasks. To our knowledge, the NSAVP dataset is the first to include stereo thermal cameras together with stereo event and monochrome cameras. The dataset and supporting software suite is available at https://umautobots.github.io/nsavp .

A Novel Algorithm for Enhancing Terrain-Aided Navigation in Autonomous Underwater Vehicles

The position error in an inertial navigation system (INS) for autonomous underwater vehicles (AUVs) increases over time. Terrain-aided navigation can assist in correcting these INS position errors. To enhance the matching accuracy under large initial position errors, an improved terrain matching algorithm comprising terrain contour matching (TERCOM), particle swarm optimization (PSO), and iterative closest contour point (ICCP), named TERCOM-PSO-ICCP, is proposed. Initially, an enhanced TERCOM with an increased rotation angle is utilized to minimize heading errors and reduce the initial position error. The similarity extremum approach evaluates the initial matching outcomes, leading to an enhanced accuracy in the initial results. Next, artificial bee colony (ABC)-optimized PSO is employed for secondary matching to further reduce the initial position error and narrow the matching area. Finally, the ICCP, using the Mahalanobis distance as the objective function, is applied for the third matching, leveraging the ICCP’s fine search capabilities. The effective combination of these three algorithms significantly improves the terrain-aided navigation matching effect. Two tests show that the improved TERCOM-PSO-ICCP effectively reduces the matching error and corrects the position of the INS.

Pluralism and the Design of Autonomous Vehicles

This paper advocates for an ethical analysis of autonomous vehicle systems (AVSs) based on a moral epistemic pluralism. This paper contends that approaching the design of intricate social technologies, such as AVSs, is most effective when acknowledging a diverse range of values. Additionally, a comprehensive ethical framework for autonomous vehicles should be applied across two interconnected layers. The first layer centers on the individual level, where each autonomous vehicle becomes a unit of moral consideration. The second layer focuses on the system level, directing moral attention toward the intricate autonomous vehicle system as a whole. Distinguishing the approach from metaphysical pluralism, the paper responds to counterarguments from a moral relativist and value monist perspectives. It concludes by emphasizing the necessity of embracing epistemic pluralism to navigate the complex ethical landscape of AVSs, urging a holistic understanding that transcends individual events and integrates system-level considerations.

Vehicle Autonomous Driving System

Autonomous vehicles or self-driving cars industry and technology have played an important role in research industry and automotive industry. Autonomous cars are those vehicles have the ability to sense there surrounding and navigate and drive themselves on roads through traffic without human interventions. In other words, they can move from one location to another one without human interaction. In this paper, an autonomous vehicle system prototype is proposed. The vehicle is able to sense its surrounding and keep going on the road on its own through traffic and other barriers such as people and traffic lights. That is, the vehicles are able to drive and detect the road signals and take decision accordingly, whether to continue or to make a turn. The proposed system uses raspberry pi microcontroller and ultrasonic sensors to detect any object, obstacle, or pedestrians in front of the vehicle and to measure the distance. In addition, a raspberry pi camera is connected to the raspberry pi to continuously take pictures of the road. These pictures will be analyzed by the raspberry pi microcontroller. The vehicle is able to reach its destination safely. A self-driving car prototype has been designed and implemented. The porotype autonomous vehicle system was tested and performed as expected.

Autonomous radar interference detection and mitigation using neural network and signal decomposition

<p>Autonomous radar interference is a challenging problem in autonomous vehicle systems. Interference signals can decrease the signal-to-interference-noise ratio (SINR), and this condition decreases the performance detection of autonomous radar. This paper exploits a neural network and signal decomposition to detect and mitigate radar interference in autonomous vehicle applications. A neural network (NN) with four inputs, one hidden layer, and one output is trained with various signal-to-noise (SNR), interference radar bandwidth, and sweep time of autonomous radar. Four inputs of NN represent SNR, mean, total harmonic distortion (THD), and root means square (RMS) of the received radar signal. Variational mode decomposition (VMD) and zeroing based on a constant false alarm rate (CFAR-Z) are used to mitigate radar interference. VMD algorithm is applied to decompose interference signals into multi-frequency sub-band. As a result, the proposed neural network can detect radar interference, and NN-VMD-CFAR-Z can increase SINR up to 2dB higher than the NN-CFAR-Z algorithm.</p>

Agricultural path detection systems using Canny-edge detection and Hough transform

Navigation is one of the crucial aspects of automation technology within the field of agriculture, such as robotics systems or autonomous agricultural vehicles. Despite many navigation systems having been developed for agricultural land, due to their high development and component costs, these systems are difficult to access for farmers or organizations with limited capital. In this study, the Canny-edge detection and Hough transform methods are implemented in a path detection system on agricultural land to find an alternative, cost-effective navigation system for autonomous farming robots or vehicles. The system is tested on ground-level view images, which are captured from a low perspective and under three different lighting conditions. The testing and experimentation process involves adjusting the parameters of the Canny-edge detection and Hough transform methods for different lighting conditions. Subsequently, an evaluation is conducted using Intersection over Union to obtain the best accuracy results, followed by fine-tuning of the canny-edge detection and Hough transform method parameters. The identified parameters, specifically a 15×15 Gaussian kernel, low threshold of 50, high threshold of 150, Hough threshold, minimum line length of 150, and maximum line gap, have been discerned as optimal for the canny-edge and Hough transform algorithms under medium lighting conditions (G=1.0). The observed efficacy of these parameter configurations suggests the method’s viability for implementation in path detection systems for agricultural vehicles or robots. This underscores its potential to deliver reliable performance and navigate seamlessly across diverse lighting scenarios within the agricultural context.

Artificial Intelligence Workload Allocation Method for Vehicular Edge Computing

Real-time applications such as smart transportation systems require minimum response time to increase performance. Incorporating edge computing, processing units near end devices, achieving fast response time. The collaboration between edge servers and cloud servers is beneficial in achieving the lowest response time by using edge servers and high computational resources by using cloud servers. The workload allocation between edge–cloud servers is challenging, especially in a highly dynamic system with multiple factors varying over time. In this paper, the workload allocation decisions among the edge servers and cloud are considered for autonomous vehicle systems. The autonomous vehicle system generates multiple tasks belonging to different AI applications running on the vehicles. The proposed method considers allocating the tasks to edge or cloud servers. The cloud servers can be reached through a cellular network or a wireless network. The proposed method is based on designing a neural network model and using a high number of features that contribute to the decision-making process. A huge dataset has also been generated for the implementation. The EdgeCloudSim is used as a simulator for implementation. The competitor's methods considered for the comparison are random, simple moving average (SMA) based, multi-armed bandit (MAB) theory-based, game theory-based, and machine learning-based workload allocation methods. The result shows an improvement in the average Quality of Experience (QoE), ranging from 8.33% to 28.57%, while the average failure rate achieved enhancement up to 50%.

Three-Dimensional Posture Estimation of Vehicle Occupants Using Depth and Infrared Images.

Pose estimation is crucial for ensuring passenger safety and better user experiences in semi- and fully autonomous vehicles. Traditional methods relying on pose estimation from regular color images face significant challenges due to a lack of three-dimensional (3D) information and the sensitivity to occlusion and lighting conditions. Depth images, which are invariant to lighting issues and provide 3D information about the scene, offer a promising alternative. However, there is a lack of strong work in 3D pose estimation from such images due to the time-consuming process of annotating depth images with 3D postures. In this paper, we present a novel approach to 3D human posture estimation using depth and infrared (IR) images. Our method leverages a three-stage fine-tuning process involving simulation data, approximated data, and a limited set of manually annotated samples. This approach allows us to effectively train a model capable of accurate 3D pose estimation with a median error of under 10 cm across all joints, using fewer than 100 manually annotated samples. To the best of our knowledge, this is the first work focusing on vehicle occupant posture detection utilizing only depth and IR data. Our results demonstrate the feasibility and efficacy of this approach, paving the way for enhanced passenger safety in autonomous vehicle systems.

Adaptive event‐triggered lateral control for autonomous vehicle system under stochastic‐sampling subject to dynamic quantization

Adaptive event‐triggered lateral control for autonomous vehicle system under stochastic‐sampling subject to dynamic quantization

Abnormal Driving Area Detection Using Multiple Vehicle Dynamic Model-Based Filter: Design and Experimental Validation

The main concern of remote control systems for autonomous ground vehicles (AGVs) is to perform the given mission according to the purpose of the operator. Although most remote systems are composed of a screen-based architecture, they are insufficient to transfer sufficient information to the remote operator. Therefore, in this paper, we present and experimentally validate an abnormal driving area detection system using an interacting multiple model (IMM) filter for the remote control system. In the proposed IMM filter, the unknown dynamic behavior of the vehicle, which changes according to changes in the driving environment, was lumped into a parameter change of the system model. As a result, we can obtain the probability of each model representing the reliability of each model, but an index can be used to infer the current status of the AGV and the driving environment. The index can help us detect both unusual behavior of the AGV such as skidding or sliding, and areas with low-friction road conditions that are not confirmed by images from the camera sensor. Thus, the remote operator can directly decide whether to continue operating or not. The proposed method is simple but useful and meaningful for the remote operator compared to the image-only method. The overall procedure of the proposed method was experimentally validated via a multi-purpose AGV on rough unpaved proving ground. Nine abnormal driving areas were discovered on the ground. In five of these areas, vehicles consistently exhibited abnormal driving behavior. The remaining four areas were confirmed to be affected by variables such as weather conditions and vehicle tire wear.

Consensus Control of Heterogeneous Uncertain Multiple Autonomous Underwater Vehicle Recovery Systems in Scenarios of Implicit Reduced Visibility

This paper investigates consensus control in heterogeneous and uncertain multiple autonomous underwater vehicle (AUV) systems under implicit reduced visibility conditions. We address challenges such as environmental uncertainties and system nonlinearity by utilizing a unified connectivity approach to model low-visibility interactions and heterogeneous multi-AUV dynamics. Our main contributions include developing a feedback linearization model for heterogeneous multi-AUV systems that accounts for uncertainties, introducing an adaptive consensus controller based on relative positioning that effectively manages implicit visual interaction limitations and validating our strategies through stability analysis and numerical simulations. Our simulations demonstrate approximately a 60% improvement in accuracy compared to previous algorithms, highlighting the practical value of our approach in AUV recovery operations. These advancements provide a robust solution for consensus control in complex underwater environments.

Adaptive fixed‐time fault‐tolerant fuzzy control of AUVs with asymmetric output constraints

AbstractThe fast and safe formation tracking problem for multiple autonomous underwater vehicle (AUV) systems (MAUVS) with asymmetric output constraints and actuator faults is studied in this article. Actuator faults are composed of loss of effectiveness and time‐varying unknown bias faults, an adaptive fixed‐time fault‐tolerant controller (AFFTC) is designed by employing fixed‐time stable theory and fuzzy logic systems. Under the designed control algorithm, the MAUVS can be practically fixed‐time stable, the tracking errors among the follower AUVs and the virtual leader AUV can converge to a small area near the origin within the fixed time, and the settling time is unaffected by the initial state of the system. To enhance the safety of MAUVS, a novel asymmetric barrier function is utilized to constrain the trajectory tracking errors within the prescribed range. Finally, the simulation results demonstrate the effectiveness of the proposed control algorithm.

Revolutionizing logistics: The impact of autonomous vehicles on supply chain efficiency

This review paper assesses the impact of autonomous vehicles (AVs) on enhancing the efficiency of supply chains, aiming to synthesize existing research and identify future research directions. Through a comprehensive analysis of literature spanning technological, economic, and regulatory perspectives, this paper maps the current landscape of AV integration in logistics. The review finds that AVs offer substantial improvements in supply chain efficiency through continuous operation and reduced human error, leading to decreased costs and improved delivery timelines. It details the potential for up to 30% enhancement in logistical operations with AV adoption. Critical barriers such as technological challenges, regulatory concerns, and cybersecurity risks are also discussed. Conclusively, the paper argues that while AVs present a promising avenue for revolutionizing supply chain management, their widespread adoption depends on overcoming existing barriers. It calls for targeted research into robust AV systems, supportive policy frameworks, and stakeholder engagement strategies to fully realize their benefits.

Многофакторный анализ эффективности автономных транспортных средств на основе метода анализа среды функционирования

The design of robotic and unmanned vehicles includes a virtual commissioning stage that provides an analysis of the operation of objects using digital twins. Conducting a large number of experiments on models leads to the use of methods for evaluating the results obtained to choose the best design solution. A method of multifactorial anal-ysis of the efficiency of the complex of autonomous vehicles (AV) for agro-industrial purposes is proposed. Multivariate analysis is performed at the virtual commissioning stage in order to plan early maintenance and repair activities. The paper uses the method of analyzing the operating environment for a comparative assessment of various operating scenarios. The method is based on solving a complex of optimization problems of linear programming. A formal description of virtual test scenarios is proposed. The method of multifactorial analysis is implemented in the process of virtual tests using optimal purpose system models and AV simulation. A procedure for virtual AV tests has been developed, including step-by-step modeling and multivariate analysis. The sets of input and output key parameters of the AV and analyzed scenarios are determined. It is proposed to perform two consecutive efficiency assessment tasks. The first task is to compare the effectiveness of individual AV and determine the limits of their effectiveness. The solution of the second task makes it possible to evaluate the effectiveness of AV system operation scenarios. The target values of changes in the key parameters of the AV and operating scenarios are obtained, leading to an increase in the efficiency of operation and maintenance. Conducting a multifactorial analysis of the results of virtual tests allows you to formulate requirements and recommendations for the design of AV maintenance systems.

Method of Forming and Algorithm of Saving a Photo Map of the Seafloor during its Photography by an Unmanned Underwater Vehicle

The paper proposes a new method and algorithm for creating a photo map of the seabed from photographs obtained from on-board photo and video systems of autonomous uninhabited underwater vehicles during fulfillment of their various missions. This photo map is formed from photographs of the seabed by sequentially overlapping them. At the same time, data on the position of the underwater vehicle received from the on-board navigation system at the time of receipt of each photograph is used to position each photo image on the generated photo map. And to rotate and scale these photographic images, data on the orientation of the underwater vehicle in the global coordinate system and its height above the bottom, also obtained from the on-board navigation system, are used. To provide quick access to the photo map, it is proposed to store it on an on-board data storage device in the form of a tile map, widely used when creating interactive geographical maps, for example, Google Maps. However, the formation of a tile map is usually carried out on the basis of all the already received photographic images, as is done with satellite images, which cannot be applied on board underwater vehicles when data from the photo-video system is received sequentially. Therefore, a special algorithm has been developed and implemented that generates a tile map during the movement of the underwater vehicle and updates only the area of the tile map in which the photo is currently being taken. This minimizes the computational load and ensures the formation of a photo map of the bottom using standard on-board computing devices. The conducted semi-natural experiments with data obtained as a result of performing real missions using the MMT-3000 underwater vehicle confirmed the high efficiency of using the developed method and algorithm.

A comparison of multi-source power supply systems for autonomous marine vehicles: The SWAMP case study

Extensive research on zero-emissions autonomous surface vehicles has been recently carried out, aiming at increasing autonomy. Due to their small size, unmanned vehicles present unique difficulties in terms of weight and dimensions when powering the vehicle with renewable energy sources. In this paper, photovoltaic source, fuel cell and Li-ion battery multi-source configurations are proposed, demonstrating by physical layout and simulation results the feasibility of the prototype, compliant with strict constraints of the vehicle under test. The energy system model of the vehicle under test is implemented in MATLAB/Simulink. A comparison of multi-source energy system configurations is proposed and validated by simulation results in terms of endurance, weight and size. If compared with the original battery-powered vehicle, as shown by comparing simulation results, the endurance is doubled, extended up to 12 h on the most favorable day of the year. The solution even complies with payload and physical dimensions constraints.

Anti-disturbance cooperative formation containment control for multiple autonomous underwater vehicles with collision-free and actuator saturation constraints

This article investigates the formation containment control (FCC) problem of multiple autonomous underwater vehicles system (MAUVS) in the simultaneous presence of disturbances, collision avoidance between autonomous underwater vehicles (AUVs) and actuator saturation constraints. Firstly, for the purpose of enhancing the resisting disturbance capability of the MAUVS, the extended state observers are designed to estimate the total disturbance comprised of internal uncertainties and external interferences in real time. In order to hinder the occurrence of complexity explosion during the backstepping design, tracking differentiators are adopted to provide estimates for the derivative of the virtual control signals. Furthermore, the auxiliary system is constructed for reducing the influence of the actuator saturation on the MAUVS. Potential functions are integrated into the designed FCC strategy, so that the collisions between AUVs can be effectively avoided while the MAUVS achieving formation containment. Afterwards, via resorting to Lyapunov stability theory, it is demonstrated that the formation tracking errors and containment errors are uniformly ultimately bounded. Lastly, the comparative simulation results are illustrated to validate the feasibility of the presented anti-disturbance FCC scheme with collision-free and anti-saturation.

Cybersecurity in Autonomous Vehicles—Are We Ready for the Challenge?

The rapid development and deployment of autonomous vehicles (AVs) present unprecedented opportunities and challenges in the transportation sector. While AVs promise enhanced safety, efficiency, and convenience, they also introduce significant cybersecurity vulnerabilities due to their reliance on advanced electronics, connectivity, and artificial intelligence (AI). This review examines the current state of cybersecurity in autonomous vehicles, identifying major threats such as remote hacking, sensor manipulation, data breaches, and denial of service (DoS) attacks. It also explores existing countermeasures including intrusion detection systems (IDSs), encryption, over-the-air (OTA) updates, and authentication protocols. Despite these efforts, numerous challenges remain, including the complexity of AV systems, lack of standardization, latency issues, and resource constraints. This review concludes by highlighting future directions in cybersecurity research and development, emphasizing the potential of AI and machine learning, blockchain technology, industry collaboration, and legislative measures to enhance the security of autonomous vehicles.

FETR: Feature Transformer for vehicle-infrastructure cooperative 3D object detection

3D object detection plays a crucial role in the perception system of autonomous vehicles, however, the vehicle’s field of view is restricted due to obstructions from nearby vehicles and buildings. Vehicle-infrastructure cooperation can compensate for the issue of visibility, but due to discrepancies in timestamps between vehicle and infrastructure sensors as well as data transmission delays, there is typically a time asynchrony between vehicle and infrastructure data. Therefore, Feature Transformer (FETR) has been introduced, which is a vehicle-infrastructure cooperative 3D object detection model utilizing Transformer as a Feature Predictor. The Transformer Predictor is capable of predicting features of future frame based on the current frame features, efficiently addressing the problem of time asynchrony. Additionally, to enhance the precision of 3D object detection, we have introduced a plug-and-play module named Mask Feature Enhancement (MFE), MFE employs a mask to amplify the features in the object region while simultaneously diminishing the features of the surrounding environment, enlarging the difference between object features and environmental features, thereby improving the detection effect. Experimental results show that FETR attains a 68.15 BEV-mAP (IoU=0.5) on the DAIR-V2X dataset, with a 200ms latency, and the data transmission is merely 6.0×104 bytes, constituting just 4.2% of the original point cloud data, outperforming current vehicle-infrastructure cooperative models in terms of both precision and data transmission.