Evolutionary Decision-Making and Planning for Autonomous Driving Based on Safe and Rational Exploration and Exploitation

Decision-making and motion planning are extremely important in autonomous driving to ensure safe driving in a real-world environment.This study proposes an online evolutionary decision-making and motion planning framework for autonomous driving based on a hybrid data-and model-driven method.First,a data-driven decision-making module based on deep reinforcement learning(DRL)is developed to pursue a rational driving performance as much as possible.Then,model predictive control(MPC)is employed to execute both longitudinal and lateral motion planning tasks.Multiple constraints are defined according to the vehicle’s physical limit to meet the driving task requirements.Finally,two principles of safety and rationality for the self-evolution of autonomous driving are proposed.A motion envelope is established and embedded into a rational exploration and exploitation scheme,which filters out unreasonable experiences by masking unsafe actions so as to collect high-quality training data for the DRL agent.Experiments with a high-fidelity vehicle model and MATLAB/Simulink co-simulation environment are conducted,and the results show that the proposed online-evolution framework is able to generate safer,more rational,and more efficient driving action in a real-world environment.

A Survey on an Emerging Safety Challenge for Autonomous Vehicles:Safety of the Intended Functionality

As the complexity of autonomous vehicles(AVs)continues to increase and artificial intelligence algorithms are becoming increasingly ubiquitous,a novel safety concern known as the safety of the intended functionality(SOTIF)has emerged,presenting significant challenges to the widespread deployment of AVs.SOTIF focuses on issues arising from the functional insufficiencies of the AVs’intended functionality or its implementation,apart from conventional safety considerations.From the systems engineering standpoint,this study offers a comprehensive exploration of the SOTIF landscape by reviewing academic research,practical activities,challenges,and perspectives across the development,verification,validation,and operation phases.Academic research encompasses system-level SOTIF studies and algorithm-related SOTIF issues and solutions.Moreover,it encapsulates practical SOTIF activities undertaken by corporations,government entities,and academic institutions spanning international and Chinese contexts,focusing on the overarching methodologies and practices in different phases.Finally,the paper presents future challenges and outlook pertaining to the development,verification,validation,and operation phases,motivating stakeholders to address the remaining obstacles and challenges.

Pavement Cracks Coupled With Shadows:A New Shadow-Crack Dataset and A Shadow-Removal-Oriented Crack Detection Approach

Automatic pavement crack detection is a critical task for maintaining the pavement stability and driving safety.The task is challenging because the shadows on the pavement may have similar intensity with the crack,which interfere with the crack detection performance.Till to the present,there still lacks efficient algorithm models and training datasets to deal with the interference brought by the shadows.To fill in the gap,we made several contributions as follows.First,we proposed a new pavement shadow and crack dataset,which contains a variety of shadow and pavement pixel size combinations.It also covers all common cracks(linear cracks and network cracks),placing higher demands on crack detection methods.Second,we designed a two-step shadow-removal-oriented crack detection approach:SROCD,which improves the performance of the algorithm by first removing the shadow and then detecting it.In addition to shadows,the method can cope with other noise disturbances.Third,we explored the mechanism of how shadows affect crack detection.Based on this mechanism,we propose a data augmentation method based on the difference in brightness values,which can adapt to brightness changes caused by seasonal and weather changes.Finally,we introduced a residual feature augmentation algorithm to detect small cracks that can predict sudden disasters,and the algorithm improves the performance of the model overall.We compare our method with the state-of-the-art methods on existing pavement crack datasets and the shadow-crack dataset,and the experimental results demonstrate the superiority of our method.

End-to-End Joint Multi-Object Detection and Tracking for Intelligent Transportation Systems

Environment perception is one of the most critical technology of intelligent transportation systems(ITS).Motion interaction between multiple vehicles in ITS makes it important to perform multi-object tracking(MOT).However,most existing MOT algorithms follow the tracking-by-detection framework,which separates detection and tracking into two independent segments and limit the global efciency.Recently,a few algorithms have combined feature extraction into one network;however,the tracking portion continues to rely on data association,and requires com‑plex post-processing for life cycle management.Those methods do not combine detection and tracking efciently.This paper presents a novel network to realize joint multi-object detection and tracking in an end-to-end manner for ITS,named as global correlation network(GCNet).Unlike most object detection methods,GCNet introduces a global correlation layer for regression of absolute size and coordinates of bounding boxes,instead of ofsetting predictions.The pipeline of detection and tracking in GCNet is conceptually simple,and does not require compli‑cated tracking strategies such as non-maximum suppression and data association.GCNet was evaluated on a multivehicle tracking dataset,UA-DETRAC,demonstrating promising performance compared to state-of-the-art detectors and trackers.

Parallel Driving in CPSS:A Unified Approach for Transport Automation and Vehicle Intelligence

The emerging development of connected and automated vehicles imposes a significant challenge on current vehicle control and transportation systems. This paper proposes a novel unified approach, Parallel Driving, a cloud-based cyberphysical-social systems(CPSS) framework aiming at synergizing connected automated driving. This study first introduces the CPSS and ACP-based intelligent machine systems. Then the parallel driving is proposed in the cyber-physical-social space,considering interactions among vehicles, human drivers, and information. Within the framework, parallel testing, parallel learning and parallel reinforcement learning are developed and concisely reviewed. Development on intelligent horizon(iHorizon)and its applications are also presented towards parallel horizon.The proposed parallel driving offers an ample solution for achieving a smooth, safe and efficient cooperation among connected automated vehicles with different levels of automation in future road transportation systems.

Analysis of Autopilot Disengagements Occurring During Autonomous Vehicle Testing

In present-day highly-automated vehicles, there are occasions when the driving system disengages and the human driver is required to take-over. This is of great importance to a vehicle's safety and ride comfort. In the U.S state of California, the Autonomous Vehicle Testing Regulations require every manufacturer testing autonomous vehicles on public roads to submit an annual report summarizing the disengagements of the technology experienced during testing. On 1 January 2016,seven manufacturers submitted their first disengagement reports:Bosch, Delphi, Google, Nissan, Mercedes-Benz, Volkswagen, and Tesla Motors. This work analyses the data from these disengagement reports with the aim of gaining abetter understanding of the situations in which a driver is required to takeover, as this is potentially useful in improving the Society of Automotive Engineers(SAE) Level 2 and Level 3 automation technologies.Disengagement events from testing are classified into different groups based on attributes and the causes of disengagement are investigated and compared in detail. The mechanisms and time taken for take-over transition occurred in disengagements are studied. Finally, recommendations for OEMs, manufacturers, and government organizations are also discussed.

Advances in Vision-Based Lane Detection：Algorithms,Integration,Assessment,and Perspectives on ACP-Based Parallel Vision

Lane detection is a fundamental aspect of most current advanced driver assistance systems(ADASs). A large number of existing results focus on the study of vision-based lane detection methods due to the extensive knowledge background and the low-cost of camera devices. In this paper, previous visionbased lane detection studies are reviewed in terms of three aspects, which are lane detection algorithms, integration, and evaluation methods. Next, considering the inevitable limitations that exist in the camera-based lane detection system, the system integration methodologies for constructing more robust detection systems are reviewed and analyzed. The integration methods are further divided into three levels, namely, algorithm, system,and sensor. Algorithm level combines different lane detection algorithms while system level integrates other object detection systems to comprehensively detect lane positions. Sensor level uses multi-modal sensors to build a robust lane recognition system. In view of the complexity of evaluating the detection system, and the lack of common evaluation procedure and uniform metrics in past studies, the existing evaluation methods and metrics are analyzed and classified to propose a better evaluation of the lane detection system. Next, a comparison of representative studies is performed. Finally, a discussion on the limitations of current lane detection systems and the future developing trends toward an Artificial Society, Computational experiment-based parallel lane detection framework is proposed.

Decision-Making in Driver-Automation Shared Control:A Review and Perspectives

Shared control schemes allow a human driver to work with an automated driving agent in driver-vehicle systems while retaining the driver’s abilities to control.The human driver,as an essential agent in the driver-vehicle shared control systems,should be precisely modeled regarding their cognitive processes,control strategies,and decision-making processes.The interactive strategy design between drivers and automated driving agents brings an excellent challenge for human-centric driver assistance systems due to the inherent characteristics of humans.Many open-ended questions arise,such as what proper role of human drivers should act in a shared control scheme?How to make an intelligent decision capable of balancing the benefits of agents in shared control systems?Due to the advent of these attentions and questions,it is desirable to present a survey on the decision making between human drivers and highly automated vehicles,to understand their architectures,human driver modeling,and interaction strategies under the driver-vehicle shared schemes.Finally,we give a further discussion on the key future challenges and opportunities.They are likely to shape new potential research directions.

Vehicle Dynamic State Estimation： State of the Art Schemes and Perspectives

Next-generation vehicle control and future autonomous driving require further advances in vehicle dynamic state estimation. This article provides a concise review, along with the perspectives, of the recent developments in the estimation of vehicle dynamic states. The definitions used in vehicle dynamic state estimation are first introduced, and alternative estimation structures are presented. Then, the sensor configuration schemes used to estimate vehicle velocity, sideslip angle, yaw rate and roll angle are presented. The vehicle models used for vehicle dynamic state estimation are further summarized, and representative estimation approaches are discussed. Future concerns and perspectives for vehicle dynamic state estimation are also discussed.

FISS GAN:A Generative Adversarial Network for Foggy Image Semantic Segmentation

Because pixel values of foggy images are irregularly higher than those of images captured in normal weather(clear images),it is difficult to extract and express their texture.No method has previously been developed to directly explore the relationship between foggy images and semantic segmentation images.We investigated this relationship and propose a generative adversarial network(GAN)for foggy image semantic segmentation(FISS GAN),which contains two parts:an edge GAN and a semantic segmentation GAN.The edge GAN is designed to generate edge information from foggy images to provide auxiliary information to the semantic segmentation GAN.The semantic segmentation GAN is designed to extract and express the texture of foggy images and generate semantic segmentation images.Experiments on foggy cityscapes datasets and foggy driving datasets indicated that FISS GAN achieved state-of-the-art performance.

Parallel Reinforcement Learning:A Framework and Case Study

In this paper, a new machine learning framework is developed for complex system control, called parallel reinforcement learning. To overcome data deficiency of current data-driven algorithms, a parallel system is built to improve complex learning system by self-guidance. Based on the Markov chain(MC) theory, we combine the transfer learning, predictive learning, deep learning and reinforcement learning to tackle the data and action processes and to express the knowledge. Parallel reinforcement learning framework is formulated and several case studies for real-world problems are finally introduced.

Simulation and Field Testing of Multiple Vehicles Collision Avoidance Algorithms

A global planning algorithm for intelligent vehicles is designed based on the A* algorithm, which provides intelligent vehicles with a global path towards their destinations. A distributed real-time multiple vehicle collision avoidance(MVCA)algorithm is proposed by extending the reciprocal n-body collision avoidance method. MVCA enables the intelligent vehicles to choose their destinations and control inputs independently,without needing to negotiate with each other or with the coordinator. Compared to the centralized trajectory-planning algorithm, MVCA reduces computation costs and greatly improves the robustness of the system. Because the destination of each intelligent vehicle can be regarded as private, which can be protected by MVCA, at the same time MVCA can provide a real-time trajectory planning for intelligent vehicles. Therefore,MVCA can better improve the safety of intelligent vehicles. The simulation was conducted in MATLAB, including crossroads scene simulation and circular exchange position simulation. The results show that MVCA behaves safely and reliably. The effects of latency and packet loss on MVCA are also statistically investigated through theoretically formulating broadcasting process based on one-dimensional Markov chain. The results uncover that the tolerant delay should not exceed the half of deciding cycle of trajectory planning, and shortening the sending interval could alleviate the negative effects caused by the packet loss to an extent. The cases of short delay(< 100100 ms) and low packet loss(< 5%) can bring little influence to those trajectory planning algorithms that only depend on V2 V to sense the context, but the unpredictable collision may occur if the delay and packet loss are further worsened. The MVCA was also tested by a real intelligent vehicle, the test results prove the operability of MVCA.

Study on the Driving Style Adaptive Vehicle Longitudinal Control Strategy

This paper presents a fusion control strategy of adaptive cruise control(ACC) and collision avoidance(CA),which takes into account a driver’s behavioral style. First, a questionnaire survey was performed to identify driver type, and the corresponding driving behavioral data were collected via driving simulator experiments, which served as the template data for the online identification of driver type. Then, the driveradaptive ACC/CA fusion control strategy was designed, and its effect was verified by virtual experiments. The results indicate that the proposed control strategy could achieve the fusion control of ACC and CA successfully and improve driver adaptability and comfort.

Parallel Planning:A New Motion Planning Framework for Autonomous Driving

Motion planning is one of the most significant technologies for autonomous driving. To make motion planning models able to learn from the environment and to deal with emergency situations, a new motion planning framework called as"parallel planning" is proposed in this paper. In order to generate sufficient and various training samples, artificial traffic scenes are firstly constructed based on the knowledge from the reality.A deep planning model which combines a convolutional neural network(CNN) with the Long Short-Term Memory module(LSTM) is developed to make planning decisions in an end-toend mode. This model can learn from both real and artificial traffic scenes and imitate the driving style of human drivers.Moreover, a parallel deep reinforcement learning approach is also presented to improve the robustness of planning model and reduce the error rate. To handle emergency situations, a hybrid generative model including a variational auto-encoder(VAE) and a generative adversarial network(GAN) is utilized to learn from virtual emergencies generated in artificial traffic scenes. While an autonomous vehicle is moving, the hybrid generative model generates multiple video clips in parallel, which correspond to different potential emergency scenarios. Simultaneously, the deep planning model makes planning decisions for both virtual and current real scenes. The final planning decision is determined by analysis of real observations. Leveraging the parallel planning approach, the planner is able to make rational decisions without heavy calculation burden when an emergency occurs.

Planning and Decision-making for Connected Autonomous Vehicles at Road Intersections:A Review

Planning and decision-making technology at intersections is a comprehensive research problem in intelligent transportation systems due to the uncertainties caused by a variety of traffic participants.As wireless communication advances,vehicle infrastructure integrated algorithms designed for intersection planning and decision-making have received increasing attention.In this paper,the recent studies on the planning and decision-making technologies at intersections are primarily overviewed.The general planning and decision-making approaches are presented,which include graph-based approach,prediction base approach,optimization-based approach and machine learning based approach.Since connected autonomous vehicles(CAVs)is the future direction for the automated driving area,we summarized the evolving planning and decision-making methods based on vehicle infrastructure cooperative technologies.Both four-way signalized and unsignalized intersection(s)are investigated under purely automated driving traffic and mixed traffic.The study benefit from current strategies,protocols,and simulation tools to help researchers identify the presented approaches’challenges and determine the research gaps,and several remaining possible research problems that need to be solved in the future.

A Situation-Aware Collision Avoidance Strategy for Car-Following

Abstract--In this paper, we discuss how to develop an appropriate collision avoidance strategy for car-following. This strategy aims to keep a good balance between traffic safety and efficiency while also taking into consideration the unavoidable uncertainty of position/speed perception/measurement of vehicles and other drivers. Both theoretical analysis and numerical testing results are provided to show the effectiveness of the proposed strategy. Index Terms--Collision avoidance, safety, traffic efficiency, uncertainty.

Hazard-evaluation-oriented Moving Horizon Parallel Steering Control for Driver-Automation Collaboration During Automated Driving

Prompted by emerging developments in connected and automated vehicles, parallel steering control, one aspect of parallel driving, has become highly important for intelligent vehicles for easing the burden and ensuring the safety of human drivers. This paper presents a parallel steering control framework for an intelligent vehicle using moving horizon optimization.The framework considers lateral stability, collision avoidance and actuator saturation and describes them as constraints, which can blend the operation of a human driver and a parallel steering controller effectively. Moreover, the road hazard and the steering operation error are employed to evaluate the operational hazardous of an intelligent vehicle. Under the hazard evaluation,the intelligent vehicle will be mainly operated by the human driver when the vehicle operates in a safe and stable manner.The automated steering driving objective will play an active role and regulate the steering operations of the intelligent vehicle based on the hazard evaluation. To verify the effectiveness of the proposed hazard-evaluation-oriented moving horizon parallel steering control approach, various validations are conducted, and the results are compared with a parallel steering scheme that does not consider automated driving situations. The results illustrate that the proposed parallel steering controller achieves acceptable performance under both conventional conditions and hazardous conditions.

Probabilistic Lane-Change Decision-Making and Planning for Autonomous Heavy Vehicles

To improve the safety and driving stability of the autonomous heavy truck, it is necessary to consider the differences of driving behavior and drivable trajectories between the heavy trucks and passenger cars. This study proposes a probabilistic decision-making and trajectory planning framework for the autonomous heavy trucks. Firstly, the driving decision process is divided into intention generation and feasibility evaluations, which are realized using the utility theory and risk assessment, respectively. Subsequently the driving decision is made and sent to the trajectory planning module. In order to reflect the greater risks of the truck to other surrounding vehicles, the aggressiveness index(AI) is proposed and quantified to infer the asymmetrical risk level of lane-change maneuver. In the planning stage, the lateral and roll dynamics stability domains are developed as the constraints to exclude the candidate trajectories that would cause vehicle instability. Finally, the simulation results are compared between the proposed model and the artificial potential filed model in the scenarios extracted from the naturalistic driving data. It is shown that the proposed framework can provide the human-like lane-change decisions and truck-friendly trajectories, and performs well in dynamic driving environments.

Post-Impact Motion Planning and Tracking Control for Autonomous Vehicles

There is an increasing awareness of the need to reduce traffic accidents and fatality due to vehicle collision.Post-impact hazards can be more serious as the driver may fail to maintain effective control after collisions.To avoid subsequent crash events and to stabilize the vehicle,this paper proposes a post-impact motion planning and stability control method for autonomous vehicles.An enabling motion planning method is proposed for post-impact situations by combining the polynomial curve and artificial potential field while considering obstacle avoidance.A hierarchical controller that consists of an upper and a lower controller is then developed to track the planned motion.In the upper controller,a time-varying linear quadratic regulator is presented to calculate the desired generalized forces.In the lower controller,a nonlinear-optimization-based torque allocation algorithm is proposed to optimally coordinate the actuators to realize the desired generalized forces.The proposed scheme is verified under comprehensive driving scenarios through hardware-in-loop tests.

Proximity Based Automatic Data Annotation for Autonomous Driving

The recent development in autonomous driving involves high-level computer vision and detailed road scene understanding.Today,most autonomous vehicles employ expensive high quality sensor-set such as light detection and ranging(LIDAR)and HD maps with high level annotations.In this paper,we propose a scalable and affordable data collection and annotation framework image-to-map annotation proximity(I2MAP),for affordance learning in autonomous driving applications.We provide a new driving dataset using our proposed framework for driving scene affordance learning by calibrating the data samples with available tags from online database such as open street map(OSM).Our benchmark consists of 40000 images with more than40 affordance labels under various day time and weather even with very challenging heavy snow.We implemented sample advanced driver-assistance systems(ADAS)functions by training our data with neural networks(NN)and cross-validate the results on benchmarks like KITTI and BDD100K,which indicate the effectiveness of our framework and training models.