Accurate recognition of driver behaviours is the basis for a reliable driver assistance system. This paper proposes a novel fusion framework for driver behaviour recognition that utilises the traffic scene and driver gaze information. The proposed framework is based on the graph neural network (GNN) and contains three modules, namely, the gaze analysing (GA) module, scene understanding (SU) module and the information fusion (IF) module. The GA module is used to obtain gaze images of drivers, and extract the gaze features from the images. The SU module provides trajectory predictions for surrounding vehicles, motorcycles, bicycles and other traffic participants. The GA and SU modules are parallel and the outputs of both modules are sent to the IF module that fuses the gaze and scene information using the attention mechanism and recognises the driving behaviours through a combined classifier. The proposed framework is verified on a naturalistic driving dataset. The comparative experiments with the state-of-the-art methods demonstrate that the proposed framework has superior performance for driving behaviour recognition in various situations.

@en

The surface texture of asphalt pavement has a significant effect on skid resistance performance. However, its contribution to the performance of skid resistance is non-homogeneous and subjects to local validity. There are also a few deep learning models that take into account the effective contact texture region. This paper proposes a convolutional neural network model based on the effective contact texture region, containing macro- and micro-scale awareness sub-modules. In this study, the asphalt mixture with varying gradations was designed to accurately obtain the effective contact texture region. Then, the textures were disentangled into macro- and micro-texture scales by applying the fast Fourier transform and fed into the model for training. Finally, the area of effective contact texture region was calculated, and the effective contact ratio parameter was then proposed using the triangulation algorithm. The results showed that the effective contact texture area of pavement varies by the asphalt mixture type. The effective contact ratio parameter exhibited a significant positive correlation (Pearson correlation coefficient is 0.901, R²= 0.8129) with skid resistance performance and was also influenced by key sieve aggregate content from 2.36 to 4.75 mm. The data of effective contact texture region following disentanglement significantly released the model performance (the relative error dropped to 1.81%). The model exhibited improved precision and performance, which can be utilized as an efficient, non-contact alternative method for skid resistance analysis.

@en

With the development of unmanned vehicle technology, unmanned vehicles have played a huge role in logistics transportation, emergency rescue and disaster relief, etc., so the research on unmanned vehicles is becoming more and more important. Road detection is an important part of environmental perception and an important factor in the realization of assisted driving and unmanned driving technology. High-precision road detection technology can provide important environmental information for efficient planning and reasonable decision-making of unmanned vehicles. Firstly, the technical framework of road detection is given, and the road detection process is introduced in detail. Then, the vision-based road detection algorithm is introduced. Finally, some related data sets in the field of road detection are collected, which provides new ideas and methods for road detection researchers.

@en

The autonomous vehicle is widely applied in various ground operations, in which motion planning and tracking control are becoming the key technologies to achieve autonomous driving. In order to further improve the performance of motion planning and tracking control, an efficient hierarchical framework containing motion planning and tracking control for the autonomous vehicles is constructed in this paper. Firstly, the problems of planning and control are modeled and formulated for the autonomous vehicle. Then, the logical structure of the hierarchical framework is described in detail, which contains several algorithmic improvements and logical associations. The global heuristic planning based artificial potential field method is developed to generate the real-time optimal motion sequence, and the prioritized Q-learning based forward predictive control method is proposed to further optimize the effectiveness of tracking control. The hierarchical framework is evaluated and validated by the numerical simulation, virtual driving environment simulation and real-world scenario. The results show that both the motion planning layer and the tracking control layer of the hierarchical framework perform better than other previous methods. Finally, the adaptability of the proposed framework is verified by applying another driving scenario. Furthermore, the hierarchical framework also has the ability for the real-time application.

@en

The skid-steered vehicle has the advantages of simple structure and strong maneuverability. Its formation driving can effectively improve safety, reduce energy consumption and exert its benefits, and has wide application prospects in military and civilian fields. Differential skid steering has strong horizontal and vertical coupling characteristics, so the tracking performance of the vehicle is poor. Therefore, it is of great significance to study horizontal and vertical joint control. Firstly, the mathematical model of the vehicle platoon is established to realize the formation control of skid-steered vehicles. Then, a combined horizontal and vertical control strategy for skid-steered vehicle formation is proposed, and a distributed model predictive controller is designed. Finally, simulation experiments verified that the designed method has good feasibility and stability.

@en

This paper proposes a linear quadratic controller based on particle swarm algorithm for the rear wheel control of four-wheel steering vehicle. Particle swarm optimization with fitness functions is used to optimize the coefficients of the weight matrix offline. The fuzzy rules following the controller is used if the road condition is terrible. The simulation results show that the LQR control model based on particle swarm optimization makes the trajectory tracking of the vehicle better and the side slip angle of the vehicle lower. It can be proved that the controller has positive effect on handling stability of the vehicle and safety of drivers.

@en

Intelligent vehicles have achieved a considerable development in technologies and can fulfill the basic functions of autonomous driving in a limited closed environment. However, results of actual road tests show that the current technologies of intelligent vehicles still have many limitations and their large-scale application in complex urban and off-road environments still faces many challenges. As one of the key technologies, the motion planning and control technology has basically formed a complete theoretical system and has been widely applied in engineering. However, the traditional methods still have some defects in practical application, such as the inability of understanding dynamic and complex scenes, poor adaptability for different scenes, high complexity of the model, and difficulty in parameter tuning. Due to the strong ability in knowledge representation and model fitting, machine learning methods have been widely applied in perception and navigation technology for intelligent vehicles. In order to solve the problems of generalization and applicability in traditional motion planning and control techniques, many researchers have also devoted themselves to exploring the usage of deep learning, reinforcement learning, and so on machine learning methods in motion planning and control policy for intelligent vehicles. In this paper, machine learning-based methods were reviewed for motion planning and control in intelligent vehicles, analyzing the existing policy learning methods for motion planning and control from three aspects, including basic framework, basic learning paradigms, and different planning and control methods based on learning. Finally, the research status and future development directions were summarized and prospected.

@en

A reliable multi-agent decision-making system is highly demanded for safe and efficient operations of connected and autonomous vehicles (CAVs). In order to represent the mutual effects between vehicles and model the dynamic traffic environments, this research proposes an integrated and open-source framework to realize different Graph Reinforcement Learning (GRL) methods for better decision-making in interactive driving scenarios. Firstly, an interactive driving scenario on the highway with two ramps is constructed. The vehicles in this scenario are modeled by graph representation, and features are extracted via Graph Neural Network (GNN). Secondly, several GRL approaches are implemented and compared in detail. Finally, The simulation in the SUMO platform is carried out to evaluate the performance of different G RL approaches. Results are analyzed from multiple perspectives to compare the performance of different G RL methods in intelligent transportation scenarios. Experiments show that the implementation of GNN can well model the interactions between vehicles, and the proposed framework can improve the overall performance of multi-agent decision-making. The source code of our work can be found at https://github.com/Jacklinkk/TorchGRL.@en

As one of the main elements of reinforcement learning, the design of the reward function is often not given enough attention when reinforcement learning is used in concrete applications, which leads to unsatisfactory performances. In this study, a reward function matrix is proposed for training various decision-making modes with emphasis on decision-making styles and further emphasis on incentives and punishments. Additionally, we model a traffic scene via graph model to better represent the interaction between vehicles, and adopt the graph convolutional network (GCN) to extract the features of the graph structure to help the connected autonomous vehicles perform decision-making directly. Furthermore, we combine GCN with deep Q-learning and multi-step double deep Q-learning to train four decision-making modes, which are named the graph convolutional deep Q-network (GQN) and the multi-step double graph convolutional deep Q-network (MDGQN). In the simulation, the superiority of the reward function matrix is proved by comparing it with the baseline, and evaluation metrics are proposed to verify the performance differences among decision-making modes. Results show that the trained decision-making modes can satisfy various driving requirements, including task completion rate, safety requirements, comfort level, and completion efficiency, by adjusting the weight values in the reward function matrix. Finally, the decision-making modes trained by MDGQN had better performance in an uncertain highway exit scene than those trained by GQN.

@en

To deal with the nonlinear interference caused by chassis movement and road surface undulations with the tracking and aiming of unmanned combat ground vehicles, a tracking and aiming adaptive control method for unmanned combat ground vehicles on the move based on reinforcement learning compensation is proposed. This method consists of a main controller and a compensation controller. The main controller uses the PID control algorithm combined with the current tracking error to obtain the main control quantity, and the compensation controller uses the Dueling DQN reinforcement learning network to process the current state of the combat vehicle as well as the road surface undulation information near the local planning path to obtain the compensation control quantity. Firstly, the integrated kinematics model of the unmanned combat ground vehicle is established. Then, the compensation control algorithm based on reinforcement learning is described. Finally, simulation and verification are performed in three-dimensional scenes based on the V-REP dynamic software. The experimental results show that the tracking and aiming control method based on reinforcement learning compensation has good adaptive ability for chassis movement and road surface undulations, which effectively improves the tracking/aiming accuracy and stability of unmanned combat vehicles.

@en

Real-time safety systems are crucial components of intelligent vehicles. This paper introduces a prediction-based collision risk assessment approach on highways. Given a point mass vehicle dynamics system, a stochastic forward reachable set considering two-dimensional motion with vehicle state probability distributions is firstly established. We then develop an acceleration prediction model, which provides multi-modal probabilistic acceleration distributions to propagate vehicle states. The collision probability is calculated by summing up the probabilities of the states where two vehicles spatially overlap. Simulation results show that the prediction model has superior performance in terms of vehicle motion position errors, and the proposed collision detection approach is agile and effective to identify the collision in cut-in crash events.@en

Precisely modeling interactions and accurately predicting trajectories of surrounding vehicles are essential to the decision-making and path-planning of intelligent vehicles. This paper proposes a novel framework based on ensemble learning to improve the performance of trajectory predictions in interactive scenarios. The framework is termed Interactive Ensemble Trajectory Predictor (IETP). IETP assembles interaction-aware trajectory predictors as base learners to build an ensemble learner. Firstly, each base learner in IETP observes historical trajectories of vehicles in the scene. Then each base learner handles interactions between vehicles to predict trajectories. Finally, an ensemble learner is built to predict trajectories by applying two ensemble strategies on the predictions from all base learners. Predictions generated by the ensemble learner are final outputs of IETP. In this study, three experiments using different data are conducted based on the NGSIM dataset. Experimental results show that IETP improves the predicting accuracy and decreases the variance of errors compared to base learners. In addition, IETP exceeds baseline models with 50% of the training data, indicating that IETP is data-efficient. Moreover, the implementation of IETP is publicly available at https://github.com/BIT-Jack/IETP.

@en

In the autonomous driving process, the decision-making system is mainly used to provide macro-control instructions based on the information captured by the sensing system. Learning-based algorithms have apparent advantages in information processing and understanding for an increasingly complex driving environment. To incorporate the interactive information between agents in the environment into the decision-making process, this paper proposes a generalized single-vehicle-based graph neural network reinforcement learning algorithm (SGRL algorithm). The SGRL algorithm introduces graph convolution into the traditional deep neural network (DQN) algorithm, adopts the training method for a single agent, designs a more explicit incentive reward function, and significantly improves the dimension of the action space. The SGRL algorithm is compared with the traditional DQN algorithm (NGRL) and the multi-agent training algorithm (MGRL) in the highway ramp scenario. Results show that the SGRL algorithm has outstanding advantages in network convergence, decision-making effect, and training efficiency.

@en

In urban environments, the complex and uncertain intersection scenarios are challenging for autonomous driving. To ensure safety, it is crucial to develop an adaptive decision making system that can handle the interaction with other vehicles. Manually designed model-based methods are reliable in common scenarios. But in uncertain environments, they are not reliable, so learning-based methods are proposed, especially reinforcement learning (RL) methods. However, current RL methods need retraining when the scenarios change. In other words, current RL methods cannot reuse accumulated knowledge. They forget learned knowledge when new scenarios are given. To solve this problem, we propose a hierarchical framework that can autonomously accumulate and reuse knowledge. The proposed method combines the idea of motion primitives (MPs) with hierarchical reinforcement learning (HRL). It decomposes complex problems into multiple basic subtasks to reduce the difficulty. The proposed method and other baseline methods are tested in a challenging intersection scenario based on the CARLA simulator. The intersection scenario contains three different subtasks that can reflect the complexity and uncertainty of real traffic flow. After offline learning and testing, the proposed method is proved to have the best performance among all methods.@en

The deep reinforcement learning-based energy management strategies (EMS) have become a promising solution for hybrid electric vehicles (HEVs). When driving cycles are changed, the neural network will be retrained, which is a time-consuming and laborious task. A more efficient way of choosing EMS is to combine deep reinforcement learning (DRL) with transfer learning, which can transfer knowledge of one domain to the other new domain, making the network of the new domain reach convergence values quickly. Different exploration methods of DRL, including adding action space noise and parameter space noise, are compared against each other in the transfer learning process in this work. Results indicate that the network added parameter space noise is more stable and faster convergent than the others. In conclusion, the best exploration method for transferable EMS is to add noise in the parameter space, while the combination of action space noise and parameter space noise generally performs poorly. Our code is available at https://github.com/BIT-XJY/RL-based-Transferable-EMS.git.@en

Predicting the trajectories of road agents is fundamental for self-driving cars. Trajectory prediction contains many sources of uncertainty in data and modeling. A thorough understanding of this uncertainty is crucial in a safety-critical task like auto-piloting a vehicle. We need to distinguish between the uncertainty caused by partial observability of all factors that may affect a drivers’ nearfuture decisions, the so-called aleatoric uncertainty, and the uncertainty of deploying a model in new scenarios that are possibly not present in the training set, the so-called epistemic uncertainty. In this paper we propose a new framework to systematically quantify both sources of uncertainty. Specifically, to approximate the spatial distribution of an agent’s future position, we propose a 2D histogram-based deep learning model combined with deep ensemble techniques for measuring both aleatoric and epistemic uncertainty by entropy-based quantities. The proposed Uncertainty Quantification Network (UQnet) employs a causal part to enhance its generalizability. Experiments on the INTERACTION dataset show that UQnet significantly improves the generalizability of the missing rate compared to the previous state-of-the-art. Further analysis shows that high aleatoric uncertainty cases are mainly caused by heterogeneous driving behaviors and unknown intended directions. Based on this aleatoric uncertainty component, we estimate the lower bounds of mean-square-error and final-displacementerror as indicators for the predictability of trajectories. Furthermore, we use the epistemic uncertainty to identify rare cases in the test set. Our results illustrate that domain knowledge on speed-dependent driving behavior is essential for adapting a model from low-speed to high-speed situations. Our paper contributes to motion forecasting with a new framework, that recasts this problem in terms of model generalisation, and puts forward methods to quantify the resulting uncertainty. @en

Pedestrian detection is an important branch of computer vision, and it has important applications in the fields of autonomous driving, artificial intelligence and video surveillance.With the rapid development of deep learning and the proposal of large-scale datasets, pedestrian detection has reached a new stage and achieves better performance. However, the performance of state-of-the-art methods is far behind the expectation, especially when occlusion and scale variance exist. Therefore, a lot of works focused on occlusion and scale variance have been proposed in the past few years. The purpose of this article is to make a detailed review of recent progress in pedestrian detection. Firstly, brief progress of pedestrian detection in the past two decades is summarized. Secondly, recent deep learning methods focusing on occlusion and scale variance are analyzed. Moreover, the popular datasets and evaluation methods for pedestrian detection are introduced. Finally, the development trend of pedestrian detection is prospected.

@en

This paper proposes a hierarchical path tracking control framework divided into the upper controller and the lower controller for double motors independently driven unmanned high-speed tracked vehicle. The upper layer generates rolling speed command of the dual-side tracks using model predictive control. The Euler method, the second-order Runge-Kutta method and the revised fourth-order Runge-Kutta method are applied to compare the control performance in this layer. Meanwhile, to surmount control command execution delay, two compensatory methods based on control law in time domain are proposed. The lower controller translates the tracks speed command from the upper layer into motors torque according to the tracked vehicle dynamic model. Experiments show that the root-mean-square lateral error is 0.1176m and the root-mean-square heading error is 0.7552° when applying the second-order Runge-Kutta method that has the best tracking accuracy in straight off-road roads at 10m/s. Experiments also demonstrate that the control framework has good tracking performance in high-speed tracking straight path and low-speed tracking curved path. Results show the steady condition lateral error of 70km/h pavement tracking is less than 0.2m and 10km/h right-angled soft off-road road is less than 0.6m.

@en

The hybrid electric system has good potential for unmanned tracked vehicles due to its excellent power and economy. Due to unmanned tracked vehicles have no traditional driving devices, and the driving cycle is uncertain, it brings new challenges to conventional energy management strategies. This paper proposes a novel energy management strategy for unmanned tracked vehicles based on local speed planning. The contributions are threefold. Firstly, a local speed planning algorithm is adopted for the input of driving cycle prediction to avoid the dependence of traditional vehicles on driver's operation. Secondly, a prediction model based on Convolutional Neural Networks and Long Short-Term Memory (CNN-LSTM) is proposed, which is used to process both the planned and the historical velocity series to improve the prediction accuracy. Finally, based on the prediction results, the model predictive control algorithm is used to realize the real-time optimization of energy management. The validity of the method is verified by simulation using collected data from actual field experiments of our unmanned tracked vehicle. Compared with multi-step neural networks, the prediction model based on CNN-LSTM improves the prediction accuracy by 20%. Compared with the traditional regular energy management strategy, the energy management strategy based on model predictive control reduces fuel consumption by 7%.

@en

Due to the complex and dynamic character of intersection scenarios, the autonomous driving strategy at intersections has been a difficult problem and a hot point in the research of intelligent transportation systems in recent years. This paper gives a brief summary of state-of-the-art autonomous driving strategies at intersections. Firstly, we enumerate and analyze common types of intersection scenarios, corresponding simulation platforms, as well as related datasets. Secondly, by reviewing previous studies, we have summarized characteristics of existing autonomous driving strategies and classified them into several categories. Finally, we point out problems of the existing autonomous driving strategies and put forward several valuable research outlooks.

@en