Autonomous Driving Research Articles

Towards Autonomous Driving: Technologies and Data for Vehicles-to-Everything Communication.

Autonomous systems are becoming increasingly relevant in our everyday life. The transportation field is no exception and the smart cities concept raises new tasks and challenges for the development of autonomous systems development which has been progressively researched in literature. One of the main challenges is communication between different traffic objects. For instance, a mobile robot system can work as a standalone autonomous system reacting to a static environment and avoiding obstacles to reach a target. Nevertheless, more intensive communication and decision making is needed when additional dynamic objects and other autonomous systems are present in the same working environment. Traffic is a complicated environment consisting of vehicles, pedestrians, and various infrastructure elements. To apply autonomous systems in this kind of environment it is important to integrate object localization and to guarantee functional and trustworthy communication between each element. To achieve this, various sensors, communication standards, and equipment are integrated via the application of sensor fusion and AI machine learning methods. In this work review of vehicular communication systems is presented. The main focus is the researched sensors, communication standards, devices, machine learning methods, and vehicular-related data to find existing gaps for future vehicular communication system development. In the end, discussion and conclusions are presented.

Reinforcement Learning Based Speed Control with Creep Rate Constraints for Autonomous Driving of Mining Electric Locomotives

The working environment of mining electric locomotives is wet and muddy coal mine roadway. Due to low friction between the wheel and rail and insufficient utilization of creep rate, there may be idling or slipping between the wheels and rails of mining electric locomotives. Therefore, it is necessary to control the creep rate within a reasonable range. In this paper, the autonomous control algorithm for mining electric locomotives based on improved ε-greedy is theoretically proven to be convergent and effective firstly. Secondly, after analyzing the contact state between the wheel and rail under wet and slippery road conditions, it is concluded that the value of creep rate is an important factor affecting the autonomous driving of mining electric locomotives. Therefore, the autonomous control method for mining electric locomotives based on creep control is proposed in this paper. Finally, the effectiveness of the proposed method is verified through simulation. The problem of wheel slipping and idling caused by insufficient friction of mining electric locomotives in coal mining environments is effectively suppressed. Autonomous operation of vehicles with optimal driving efficiency can be achieved through quantitative control and utilization of the creep rate between wheels and rails.

Continuity of GNSS as a critical attribute for safety applications in land transport

The Global Navigation Satellite System (GNSS) is widely used for air traffic management—more than 150,000 aircraft and 5000 airports worldwide are equipped with SBAS (Satellite-based augmentation system) technology, which contributes to safer and more efficient air operations. The next challenge is to extend GNSS positioning to maritime, autonomous cars and railway control systems preserving their safety requirements. The main parameter is the integrity of the GNSS positioning, although the time for which the integrity is guaranteed, defined by continuity, the most demanding requirement for aviation applications, has not been sufficiently investigated for land transportation. The aim of this paper is to close this gap by clarifying: (1) where the requirement for GNSS continuity comes from, (2) why GNSS continuity is needed in land transport, and (3) how GNSS-based applications can be made more reliable when needed. Using a comparative analysis, the continuity requirements in aviation, rail, maritime, and road transport have been investigated showing their importance for railways and automotive control, paving the way to eventually update the current EN 50126 (RAMS) and ISO/TR 4804 standards respectively for railways and automated cars. One of the main findings, through Markov modeling, is the improvement of the Mean Time to System Failure (MTTFsys) that for the railway safety-of-life applications can be significantly increased from about 521 h up to 5 × 105 h. These results can contribute to accelerating the adoption of GNSS positioning for automated land transportation, by exploiting the extensive experience brought by the aviation sector where GNSS was introduced 20 years ago.

Future Trends: Predicting the Evolution of Rail and Road Transportation in Logistics

The logistics industry represents the vital circulatory system of the global economy, responsible for the efficient movement of goods across sprawling supply chain networks. Historically, rail and road transportation modes have been the workhorses of logistics operations, facilitated by their extensive infrastructure and versatility in cargo handling. However, the convergence of rapid technological advancements, evolving consumer expectations, and mounting environmental pressures is catalyzing a transformative shift in the future evolution of these pivotal transportation sectors. This comprehensive study delves into the emerging trends that are poised to reshape the rail and road transportation landscape within the logistics realm. By synthesizing insights from industry leaders, academic research, and cutting-edge technological developments, the study aims to provide a holistic perspective on the forthcoming changes and their far-reaching implications. The study’s main objective is to integrate semi-autonomous and autonomous cars into fleets of transportation vehicles. The advent of self-driving trucks and trains promises to revolutionize logistics operations by enhancing efficiency, reducing human error, and optimizing route planning. Concurrently, this transition raises critical concerns related to workforce displacement, regulatory frameworks, and cybersecurity vulnerabilities that necessitate proactive mitigation strategies. Moreover, the logistics industry faces mounting pressure to embrace sustainable practices and curtail its environmental footprint. Consequently, the rail sector is anticipated to experience a renaissance fueled by its inherent energy efficiency and lower carbon emissions compared to road transportation. Innovations in electric and hydrogen-powered locomotives, coupled with advancements in rail infrastructure, could position rail as a greener and more cost-effective solution for long-haul freight movement. Concurrently, the road transportation sector is poised for a transformation driven by the proliferation of electric and alternative fuel vehicles. This adoption is projected to accelerate as battery technology improves, charging infrastructure expands, and governments implement stricter emissions regulations. Furthermore, the integration of advanced telematics systems and realtime data analytics could optimize route planning, load management, and vehicle maintenance, further enhancing the operational efficiency of road-based logistics.

Autonomous Vehicles: From Technology to Law and Regulation

The advent of autonomous vehicles marks a convergence of technological advancement, societal change, and regulatory evolution with profound implications. This transformative technology has shifted from science fiction to reality, sparking extensive discourse encompassing technological progress, legal intricacies, and ethical considerations. This scholarly exploration navigates the intricate terrain of autonomous vehicles, tracing their evolution from a technological marvel to a catalyst for profound changes in legal frameworks, regulations, and societal dynamics. Our journey begins by delving into the technological foundations and significant achievements that underpin autonomous vehicle development. This illuminates the complex algorithms, sensor arrays, and machine-learning capabilities driving these self-driving vehicles. Transitioning from technology, we delve into the intricate realm of law and regulation, where established norms of road safety, liability, and privacy intersect with autonomy's transformative potential. This paper scrutinizes the intricate network of legal frameworks and ethical considerations governing the testing, deployment, and daily operation of autonomous cars. With their integration into transportation systems gaining prominence, balancing innovation with regulation assumes critical importance. This study embraces technology, law, and regulation, aiming to understand the multifaceted dimensions arising when autonomous cars intersect with legal and ethical frameworks. We explore the dynamic nature of global regulatory paradigms, analyzing obstacles, achievements, and ethical dilemmas shaping progress. As we embark on this endeavor, our goal is to offer a valuable resource elucidating the intricate process of transitioning from technology to legal frameworks and regulations within the era of autonomous vehicles.

Civil Liability for damage caused by Autonomous Cars

Civil Liability for damage caused by Autonomous Cars

Safe Autonomous Driving with Latent Dynamics and State-Wise Constraints.

Autonomous driving has the potential to revolutionize transportation, but developing safe and reliable systems remains a significant challenge. Reinforcement learning (RL) has emerged as a promising approach for learning optimal control policies in complex driving environments. However, existing RL-based methods often suffer from low sample efficiency and lack explicit safety constraints, leading to unsafe behaviors. In this paper, we propose a novel framework for safe reinforcement learning in autonomous driving that addresses these limitations. Our approach incorporates a latent dynamic model that learns the underlying dynamics of the environment from bird's-eye view images, enabling efficient learning and reducing the risk of safety violations by generating synthetic data. Furthermore, we introduce state-wise safety constraints through a barrier function, ensuring safety at each state by encoding constraints directly into the learning process. Experimental results in the CARLA simulator demonstrate that our framework significantly outperforms baseline methods in terms of both driving performance and safety. Our work advances the development of safe and efficient autonomous driving systems by leveraging the power of reinforcement learning with explicit safety considerations.

Overtaking Mechanisms Based on Augmented Intelligence for Autonomous Driving: Data Sets, Methods, and Challenges

Overtaking Mechanisms Based on Augmented Intelligence for Autonomous Driving: Data Sets, Methods, and Challenges

Using a Diffusion Model for Pedestrian Trajectory Prediction in Semi-Open Autonomous Driving Environments

Using a Diffusion Model for Pedestrian Trajectory Prediction in Semi-Open Autonomous Driving Environments

Vehicle Detection Algorithms for Autonomous Driving: A Review.

Autonomous driving, as a pivotal technology in modern transportation, is progressively transforming the modalities of human mobility. In this domain, vehicle detection is a significant research direction that involves the intersection of multiple disciplines, including sensor technology and computer vision. In recent years, many excellent vehicle detection methods have been reported, but few studies have focused on summarizing and analyzing these algorithms. This work provides a comprehensive review of existing vehicle detection algorithms and discusses their practical applications in the field of autonomous driving. First, we provide a brief description of the tasks, evaluation metrics, and datasets for vehicle detection. Second, more than 200 classical and latest vehicle detection algorithms are summarized in detail, including those based on machine vision, LiDAR, millimeter-wave radar, and sensor fusion. Finally, this article discusses the strengths and limitations of different algorithms and sensors, and proposes future trends.

Focused Test Generation for Autonomous Driving Systems

Testing Autonomous Driving Systems (ADSs) is crucial to ensure their reliability when navigating complex environments. ADSs may exhibit unexpected behaviours when presented, during operation, with driving scenarios containing features inadequately represented in the training dataset. To address this shift from development to operation, developers must acquire new data with the newly observed features. This data can be then utilised to fine tune the ADS, so as to reach the desired level of reliability in performing driving tasks. However, the resource-intensive nature of testing ADSs requires efficient methodologies for generating targeted and diverse tests. In this work, we introduce a novel approach, DeepAtash-LR , that incorporates a surrogate model into the focused test generation process. This integration significantly improves focused testing effectiveness and applicability in resource-intensive scenarios. Experimental results show that the integration of the surrogate model is fundamental to the success of DeepAtash-LR . Our approach was able to generate an average of up to 60 × more targeted, failure-inducing inputs compared to the baseline approach. Moreover, the inputs generated by DeepAtash-LR were useful to significantly improve the quality of the original ADS through fine tuning.

Dense Sample Deep Learning.

Deep learning (DL), a variant of the neural network algorithms originally proposed in the 1980s (Rumelhart etal., 1986), has made surprising progress in artificial intelligence (AI), ranging from language translation, protein folding (Jumper et al., 2021), autonomous cars, and, more recently, human-like language models (chatbots). All that seemed intractable until very recently. Despite the growing use of DL networks, little is understood about the learning mechanisms and representations that make these networks effective across such a diverse range of applications. Part of the answer must be the huge scale of the architecture and, of course, the large scale of the data, since not much has changed since 1986. But the nature of deep learned representations remains largely unknown. Unfortunately, training sets with millions or billions of tokens have unknown combinatorics, and networks with millions or billions of hidden units can't easily be visualized and their mechanisms can't be easily revealed. In this letter, we explore these challenges with a large (1.24 million weights VGG) DL in a novel high-density sample task (five unique tokens with more than 500 exemplars per token), which allows us to more carefully follow the emergence of category structure and feature construction. We use various visualization methods for following the emergence of the classification and the development of the coupling of feature detectors and structures that provide a type of graphical bootstrapping. From these results, we harvest some basic observations of the learning dynamics of DL and propose a new theory of complex feature construction based on our results.

Use possibility analysis for wind energy in autonomous car interior heating

Introduction. When driving cars in winter on intercity routes, problems may arise if the highways are covered with snow as a result of heavy snowfalls, which can last for several days or more, and movement during this period becomes impossible. An analysis of statistical data on weather events in winter has shown that in many regions, the closure of trails as a result of prolonged snowfalls can reach up to fifteen percent. In this case, an autonomous energy source, which it receives through the use of environmental properties, can provide significant assistance in improving the situation and increasing safety. The paper considers the technical solution and carries out a theoretical analysis of its characteristics aimed at solving the problem of autonomous power supply of the car interior in severe weather conditions.Materials and methods. The paper presents a technical solution and a theoretical analysis of its characteristics aimed at solving the problem of autonomous power supply of the car interior in severe weather conditions.Results. Based on the results of the information search, the design of a compact low-power wind power plant is proposed. The calculation of the heat loss of the car interior, the characteristics of a compact wind power plant was performed and the generated power required to heat the car interior in an extreme situation was determined.Discussion and conclusions. It has been established that the presented design of a low-power wind power plant can be used as a prototype for the development of an industrial design of an autonomous energy source for heating a car in extreme conditions. Similar installations can be used in other industries, for example, tourism, agriculture, and geological exploration.

Cooperative Vehicle Infrastructure System or Autonomous Driving System? From the Perspective of Evolutionary Game Theory

In this paper, we explore the trade-offs between public and private investment in autonomous driving technologies. Utilizing an evolutionary game model, we delve into the complex interaction mechanisms between governments and auto manufacturers, focusing on how strategic decisions impact overall outcomes. Specifically, we predict that governments may opt for strategies such as constructing and maintaining infrastructure for Roadside Infrastructure-based Vehicles (RIVs) or subsidizing high-level Autonomous Driving Vehicles (ADVs) without additional road infrastructure. Manufacturers’ choices involve deciding whether to invest in RIVs or ADVs, depending on governmental policies and market conditions. Our simulation results, based on scenarios derived from existing economic data and forecasts on technology development costs, suggest that government subsidy policies need to dynamically adjust in response to manufacturers’ shifting strategies and market behavior. This dynamic adjustment is crucial as it addresses the evolving economic environment and technological advancements, ensuring that subsidies effectively incentivize the desired outcomes in autonomous vehicle development. The findings of this paper could serve as valuable decision-making tools for governments and auto manufacturers, guiding investment strategies that align with the dynamic landscape of autonomous driving technology.

Navigating Abnormal Environmental Conditions: Traffic Signal and Obstacle Detection for Autonomous Vehicles

Autonomous cars are the future smart cars anticipated to be efficient and crash avoiding ideal urban car of the future. The purpose of the project is to build up a car which can move with and without any driver. It just need a command from the user and then it will start moving towards the destination. The project aims to build a autonomous car prototype using Raspberry Pi processor. Sensor interfacing is used to provide necessary data from the real world to the car. The car is capable of reaching the given destination safely and intelligently thus avoiding the risk of human errors. Many existing algorithms like traffic light detection, obstacle detection and speed control are combined together to provide the necessary control to the car. The technology that autonomous cars and “connected vehicles” use would allow the vehicles to communicate with roadside infrastructure like traffic lights and road congestion, and then use this information to curtail fuel consumption and emissions significantly.

Visualization and Visual Analytics in Autonomous Driving.

Autonomous driving is no longer a topic of science fiction. Advancements of autonomous driving technologies are now reliable. Effectively harnessing the information is essential for enhancing the safety, reliability, and efficiency of autonomous vehicles. In this article, we explore the pivotal role of visualization and visual analytics (VA) techniques used in autonomous driving. By employing sophisticated data visualization methods, VA, researchers, and practitioners transform intricate datasets into intuitive visual representations, providing valuable insights for decision-making processes. This article delves into various visualization approaches, including spatial-temporal mapping, interactive dashboards, and machine learning-driven analytics, tailored specifically for autonomous driving scenarios. Furthermore, it investigates the integration of real-time sensor data, sensor coordination with VA, and machine learning algorithms to create comprehensive visualizations. This research advocates for the pivotal role of visualization and VA in shaping the future of autonomous driving systems, fostering innovation, and ensuring the safe integration of self-driving vehicles.

Efficient Baselines for Motion Prediction in Autonomous Driving

Efficient Baselines for Motion Prediction in Autonomous Driving

Safety-Aware Causal Representation for Trustworthy Offline Reinforcement Learning in Autonomous Driving

Safety-Aware Causal Representation for Trustworthy Offline Reinforcement Learning in Autonomous Driving

An Incentive Mechanism for Long-Term Federated Learning in Autonomous Driving

An Incentive Mechanism for Long-Term Federated Learning in Autonomous Driving

Improving the Robustness of Pedestrian Detection in Autonomous Driving With Generative Data Augmentation

Improving the Robustness of Pedestrian Detection in Autonomous Driving With Generative Data Augmentation