Search Results

Roundabouts are one of the most complex traffic environments in urban roads, and a key challenge for intelligent driving decision-making. Deep reinforcement learning, as an emerging solution for intelligent driving decisions, has the advantage of avoiding complex algorithm design and sustainable iteration. For the decision difficulty in roundabout scenarios, this paper proposes an artificial potential field based Soft Actor-Critic (APF-SAC) algorithm. Firstly, based on the Carla simulator and Gym framework, a reinforcement learning simulation system for roundabout driving is built. Secondly, to reduce reinforcement learning exploration difficulty, global path planning and path smoothing algorithms are designed to generate and optimize the path to guide the agent.

An autonomous vehicle is a comprehensive intelligent system that includes environment sensing, vehicle localization, path planning and decision-making control, of which environment sensing technology is a prerequisite for realizing autonomous driving. In the early days, vehicles sensed the surrounding environment through sensors such as cameras, radar, and lidar. With the development of 5G technology and the Vehicle-to-everything (V2X), other information from the roadside can also be received by vehicles. Such as traffic jam ahead, construction road occupation, school area, current traffic density, crowd density, etc. Such information can help the autonomous driving system understand the current driving environment more clearly. Vehicles are no longer limited to areas that can be sensed by sensors. Vehicles with different autonomous driving levels have different adaptability to the environment.

Electro-Mechanical Braking (EMB) system, which completely abandons the traditional hydraulic device, realizes complete human-vehicle decoupling and integrates various functions without adding additional accessories, could meet the requirements of the future intelligent driving technology for high-quality braking control. However, there are significant internal interference of nonlinear characteristics such as mechanical friction and system variable stiffness during the actual working process of EMB, and these make the accuracy and rate of the clamping force control decline. This paper proposes a precise clamping force control strategy for EMB based on nonlinear characteristics compensation. First, we systematically analyze the working principle of EMB, and establish the mathematical model of EMB system including motor, transmission mechanism and friction. At the same time, some typical experiments are designed to identify internal parameters of friction model.

The Electronic Mechanical Braking (EMB) system, which offers advantages such as no liquid medium and complete decoupling, can meet the high-quality active braking and high-intensity regenerative braking demands proposed by intelligent vehicles and is considered one of the ideal platforms for future chassis. However, traditional control strategies with fixed clamping force tracking parameters struggle to maintain high-quality braking performance of EMB under variable braking requests, and the nonlinear friction between mechanical components also affects the accuracy of clamping force control. Therefore, this paper presents an adaptive clamping force control strategy for the EMB system, taking into account the resistance of nonlinear friction. First, an EMB model is established as the simulation and control object, which includes the motor model, transmission model, torque balance model, stiffness model, and friction model.

In the field of autonomous driving trajectory planning, it’s virtual to ensure real-time planning while guaranteeing feasibility and robustness. Current widely adopted approaches include decoupling path planning and velocity planning based on optimization method, which can’t always yield optimal solutions, especially in complex dynamic scenarios. Furthermore, search-based and sampling-based solutions encounter limitations due to their low resolution and high computational costs. This paper presents a novel spatio-temporal trajectory planning approach that integrates both search-based planning and optimization-based planning method. This approach retains the advantages of search-based method, allowing for the identification of a global optimal solution through search. To address the challenge posed by the non-convex nature of the original solution space, we introduce a spatio-temporal semantic corridor structure, which constructs a convex feasible set for the problem.

Autonomous driving in real-world urban traffic must cope with dynamic environments. This presents a challenging decision-making problem, e.g. deciding when to perform an overtaking maneuver or how to safely merge into traffic. The traditional autonomous driving algorithm framework decouples prediction and decision-making, which means that the decision-making and planning tasks will be carried out after the prediction task is over. The disadvantage of this approach is that it does not consider the possible impact of ego vehicle decisions on the future states of other agents. In this article, a decision-making and planning method which considers longitudinal interaction is represented. The method’s architecture is mainly composed of the following parts: trajectory sampling, forward simulation, trajectory scoring and trajectory selection. For trajectory sampling, a lattice planner is used to sample three-dimensionally in both the time horizon and the space horizon.

Subframe is an important part of automobile chassis, which is connected with body, suspension control arm, powertrain mount, etc. The dynamic stiffness value of the connection point is an important performance index of the subframe, which affects the vibration of the vehicle body. This paper introduces the basic concept and related theory of dynamic stiffness, derives the theoretical formula of dynamic stiffness, and analyzes the frequency response of the key points of the subframe. In view of the fact that the dynamic stiffness of the subframe of a certain vehicle model is not up to the standard at some connection points, the dynamic stiffness CAE simulation analysis is carried out to determine the frequency range of insufficient dynamic stiffness and the connection points that need to be optimized.

Lane detection and tracking play a key role in autonomous driving, not only in the LKA System but help estimate the pose of the vehicle. While there has been significant development in recent years, traditional outdoor SLAM algorithms still struggle to provide reliable information in challenging dynamic environments such as lack of roadside landscape or surrounding vehicles at almost the same speed or on the road in the woods. On the structured road, lane markings as static semantic features may provide a stable landmark assist in robust localization. As most of the current lane detection work mainly on separated images ignoring the relationship between adjacent frames, we propose a pixel-level lane tracking method for autonomous vehicles. In this paper, we introduce a deep network to detect and track lane features. The network has two parallel branches. One branch detects the lane position, while the other extracts the point description on a pixel level.

Autonomous driving technology, as the product of the fifth stage of the information technology revolution, is of great significance for improving urban traffic and environmentally friendly sustainable development. Autonomous driving can be divided into three main modules. The input of the decision module is the perception information from the perception module and the output of the control strategy to the control module. The deep reinforcement learning method proposes an end-to-end decision-making system design scheme. This paper adopts the Deep Deterministic Policy Gradient Algorithm (DDPG) that incorporates the Priority Experience Playback (PER) method. The framework of the algorithm is based on the actor-critic network structure model. The model takes the continuously acquired perception information as input and the continuous control of the vehicle as output.

To efficiently reduce the required experimental data and improve the prediction accuracy, a virtual engine model has been built by integrating an artificial neural network (ANN) system consisting of multiple subnets with the genetic algorithm (GA). The GA algorithm could reduce the risk of local minima and lead to a more efficient training process. The engine model has been adopted to predict the combustion phases (including CA10, CA50 and CA90), exhaust gas temperature, brake specific fuel consumption rate (be) and engine emissions which are un-burnt hydrocarbon (UBHC), NOx and CO. The results are then compared with the experimental data from around 5000 operating points of a boosted DISI engine running at universal performance map and conditions with various valve timing configurations. The mean absolute errors of combustion phases are all below 1.0 crank angle degree. The averaged errors of the exhaust gas temperature and be are 10.1 K and 1.1%, respectively.

Adaptive Cruise Control (ACC) includes three modes: cruise control, car following control, and autonomous emergency braking. Among them, the car following control mode is mainly used to manage the speed and vehicle spacing approach the preceding vehicle within the range of smooth acceleration changes. In addition, although the motion information signal of the preceding vehicle can be collected by auxiliary equipment, it is still a random variable and normally regarded as a disturbance to affect the performance of vehicle controller. Therefore, this paper proposed an ACC strategy considering the disturbance of the preceding vehicle and multi-objective optimization.

Effectively detecting road boundaries in real time is critical to the applications of autonomous vehicles, such as vehicle localization, path planning and environmental understanding. To precisely extract the road boundaries from the 3D-LiDAR data, a dedicated algorithm consisting of four steps is proposed in this paper. The steps are as follows: Firstly, the 3D-LiDAR data is pre-processed by employing the RANSAC method, the ground points are quickly separated from the original 3D-LiDAR point cloud to reduce the disturbance from the obstacles on the road, this greatly decreases the size of the point cloud to be processed. Secondly, based on the principle of 3D-LiDAR scanning, the ground points are divided into scan layers. And the road boundary points of each scan layer are detected by using three spatial features based on sliding window.

In the field of automatic driving, the combination of 3D LIDAR and inertial measurement unit (IMU) is a common sensor configuration scheme in laser point-cloud localization, high-precision map making and point-cloud target detection. So it is critical to calibrate LIDAR and IMU accurately. At present, due to the large volume and high cost of 3D LIDAR with high-line-number(Such as 64 lines or 128 lines), the configuration scheme of using multiple low-line-number 3D LIDARs appears in the automatic driving vehicle sensing system. However, the common calibration methods are not suitable for multi 3D LIDARs and IMU parameters calibration on autonomous vehicle, which have the disadvantages of cumbersome implementation and low accuracy. In this paper, a joint calibration test platform composed of dual LIDARs and IMU is assembled, and a method of precise automatic calibration based on GPS/RTK data is proposed.

It’s hard to achieve complete self-driving using hand-crafting generalized decision-making rules, while the end-to-end self-driving system is low in complexity, does not require hand-crafting rules, and can deal with complex situations. Modular-based self-driving systems require multi-task fusion and high-precision maps, resulting in high system complexity and increased costs. In end-to-end self-driving, we usually only use camera to obtain scene status information, so image processing is very important. Numerous deep learning applications benefit from multi-task learning, as the multi-task learning can accelerate model training and improve accuracy with combine all tasks into one model, which reduces the amount of calculation and allows these systems to run in real-time. Therefore, the approach of obtaining rich scene state information based on multi-task learning is very attractive. In this paper, we propose an approach to multi-task learning for semantics, geometry and motion.

In order to solve such problems that the millimeter-wave radar is of large computation, poor robustness and low precision of the target tracking algorithm, this paper presents an algorithmic framework for millimeter-wave radar tracking of target-vehicles. The target measurement information outside the millimeter- wave radar detection range is eliminated by the data plausibility judgment method based on the millimeter-wave radar detection parameters. Target clustering is made using Manhattan distance, to eliminate clutter interference and cluster multiple target measurements into one. The data association is made by use of nearest neighbor to determine the correspondence between information received measured by the radar and the real target. The vehicle is the key detection target of the vehicle millimeter-wave radar during road driving.

The high popularity of automobiles has led to frequent collisions. According to the latest statistics of the United Nations, about 1.25 million people worldwide die from road traffic accidents each year. In order to improve the safety of vehicles in driving, the active safety system has become a research hotspot of various car companies and research institutions around the world. Among them, the more mature and popular active security system are Forward Collision Warning(FCW) and Autonomous Emergency Braking(AEB). However, the current active safety system is based on traditional sensors such as radar and camera. Therefore, the system itself has many limitations due to the shortage of traditional sensors. Compared to traditional sensors, Vehicle to Everything (V2X) technology has the advantages of richer vehicle parameter information, no perceived blind spots, dynamic prediction of dangerous vehicle status, and no occlusion restriction.

GAC Group has set up two modular engine families, G and GS, for various vehicle classes equipping demands. G family engines, which have already gone through three generations, target for the higher torque and power, the lower fuel consumption and the future strict emission standards. For the latest generation, new technologies were added to achieve the development goals based on the previous modular engines. For example, miller combustion cycle with increased compression ratio is introduced in the newer engine combustion system. Additional key technologies such as 350 bar injection system and high tumble intake ports are also applied. The combustion system development, which established on the GAC Combustion Controlling System (GCCS), was facilitated by integrated use of advanced optical measurements and computational fluid dynamics for improving the in-cylinder flow, fuel sprays and the interaction between them.

With the increasing number of urban cars, parking has become the primary problem that people face in daily life. Therefore, many scholars have studied the automatic parking system. In the existing research, most of the path planning methods use the combined path of arc and straight line. In this method, the path curvature is not continuous, which indirectly leads to the low accuracy of path tracking. The parking path designed using the fifth-order polynomial is continuous, but its curvature is too large to meet the steering constraints in some cases. In this paper, a continuous-curvature parking path is proposed. The parking path tracker based on Model Predictive Control (MPC) algorithm is designed under the constraints of the control accuracy and vehicle steering. Firstly, in order to make the curvature of the parking path continuous, this paper superimposes the fifth-order polynomial with the sigmoid function, and the curve obtained has the continuous and relatively small curvature.

Traffic accidents avoidance is one of the main advantages for automated vehicles. As one of the main causes of vehicle collision accidents, lane changing of the ego vehicle in case that the obstacle vehicles appear in the blind spot with uncertain motion intentions is one of the main goals for the automated vehicle. An intention-aware lane changing collision assistance strategy basing on traffic situation assessment in the complex traffic scenarios is proposed in this paper. Typical Regions of Interest (ROI) within the detection range of the blind spots are selected basing on the road topology structures and state space consisting of the ego vehicle and the obstacle vehicles. Then the motion intentions of the obstacle vehicles in ROI are identified basing on Gaussian Mixture Models (GMM) and the corresponding motion trajectories are predicted basing on the state equation.

The Electromechanical Brake Booster system (EMBB) integrates active braking and energy recovery and becomes a novel brake-by-wire solution that substitutes the vacuum booster. While the intelligent unmanned vehicle is in unstable state, the EMBB can improve the vehicle yaw stability more quickly and safely. In this paper, a new type of integrated EMBB has been designed, which mainly includes two parts: servo motor unit and hydraulic control unit. Aiming at the dynamic instability problem of intelligent unmanned vehicle, a three-layer vehicle yaw stability control structure including decision layer, distribution layer and execution layer is proposed based on integrated EMBB. Firstly, the decision layer calculates the ideal yaw rate and the side slip angle of the vehicle with the classic 2DOF vehicle dynamics model. The boundary of the stable region is determined by the phase plane method and the additional yaw moment is determined by the feedback PI control algorithm.