Proceedings of China SAE Congress 2023: Selected Papers

Inhaltsverzeichnis

1. Frontmatter

2. An Improved GA-Based Method for Generating Critical Collision Scenarios of AEB on Longitudinal Slope

   Man Zhang, Siyuan He, Jize Wen, Wendong Cheng

   Abstract

   To accelerate the testing of the function boundary of AEB control strategy on a longitudinal slope, this paper proposed an improved GA-based method for generating critical collision scenarios of AEB on the longitudinal slope. The range constraint of AEB scenario parameters, the longitudinal dynamic constraint of slope and the emergency braking demand constraint of longitudinal slope were analyzed. Furthermore, by examining the collision conditions of AEB when the front vehicle is in three states (parking or maintaining a constant speed, accelerating and decelerating), an objective function for AEB critical collision scenarios with a slope was designed based on the satisfaction criterion model. The reasonable range of each parameter was determined by assessing the degree of constraint between AEB scenario parameters. An iterative operator was then designed to satisfy the constraints of the scene parameters and eliminate invalid scenarios. The results demonstrate that, under identical conditions, the proposed method and the improved GA method with penalty term generated effective scenarios in an average time of 0.45 s and 0.91 s, respectively. This indicates that the proposed method has nearly doubled the generation efficiency. By analyzing the critical collision scenario areas at coverage ratio of 50%, 80% and 100%, it was observed that the scenario area density increased with higher coverage ratio. The critical collision scene domain of the leading vehicle exhibited a distinct boundary under different conditions, with the influence of various longitudinal slopes on the critical collision scenario domain being significant. When employing the AEB function with fixed and adaptive time-to-collision strategies, the AEB systems were activated. In some critical scenarios, the former system resulted in a collision while the latter system successfully avoided a collision. This effectively tests the functional boundary of the AEB control strategy.

3. Application of Human-Machine Co-driving in Multi-vehicle Formation of Deepracer Automatic Driving Platform

   Heng Ye, Yanbo Liu, Jiahao Xu, Haikuo Du, Weiqi Sun, Wenchao Xu, Zhengyu Li, Zanwei Shen, Yan Liu

   Abstract

   This article introduces the application of human-machine collaborative driving in multi-car platooning based on the Amazon DeepRacer autonomous driving platform, which includes the use of AirTag recognition and lane recognition. In traditional autonomous driving, vehicle control is mostly managed by algorithms with little human intervention. In our research, we enable both algorithms and drivers to share control, with algorithms providing precision and drivers offering adaptability. We discuss the principles and implementation methods of AirTag recognition and how the location information obtained from AirTag recognition can be applied to multi-car platooning. Furthermore, we explore the relevant technologies of lane recognition, which is the core of our autonomous driving system. Finally, we propose some improvement measures for practical application to ensure the safety and reliability of human-machine collaborative driving in multi-car platooning. The results of this article shows that manual driving can handle special situations, such as changing lanes, avoiding complex obstacles, and so on. Autonomous driving is more suitable for regular road driving and emergency situations.

4. Post-Processing Method for Lane Detection Based on Prior Knowledge and Distance Penalty

   Shang Jiang, Zhishuo Hu, Yuan Wang, Bofu Wu

   Abstract

   Applying post-processing techniques to extract lane instances from binary segmentation images is an important step in lane detection based on semantic segmentation. This paper proposes a lane instance extraction algorithm based on prior knowledge and distance penalty modules, improving the random sample consensus algorithm. The prior knowledge module uses existing lane prior knowledge to set filtering conditions in the algorithm iteration process, enhancing the robustness of instance extraction. The distance penalty module applies an evaluation penalty based on the distance of lane feature points, solving the deviation at the near end of the lane model caused by a sudden change in the number of feature points due to perspective transformation and improving the accuracy of instance extraction. The experimental results on a large amount of testing data from various traffic scenarios demonstrate that the proposed algorithm can be applied to multi-lane detection post-processing tasks containing any number of lanes, with an instance extraction accuracy rate of 98.53%. It has good instance extraction capabilities for lanes with any curvature, with an average extraction error as low as 16.85 pixels, and stable extraction performance for various types of lanes, such as solid and dashed lines.

5. Vehicle Platoon Motion Planning Based on Dual-Mode Strategy

   Ying Lv, Liangliang Sun, Haomiao Li, Hongyu Hu, Lei He

   Abstract

   Intelligent connected vehicle platooning technology utilizes vehicle-to-vehicle communication to enhance urban road efficiency and ensure driving safety. This paper focuses on the vehicle platoon motion planning algorithm, which constructs a reference path for platoon vehicles based on smoothed historical trajectory points of the lead vehicle. The algorithm generates path boundaries considering road and obstacle information, creates candidate paths using lateral and longitudinal sampling, and selects the optimal path using a cost function. A velocity planning algorithm is developed for platoon vehicles, incorporating dual-mode strategies for platoon cruise and obstacle avoidance based on different scenarios. A safety risk assessment is established, guiding the choice of speed planning mode. Real-world experiments validate the planning outcomes, evaluating speed, acceleration, and inter-vehicle distance. Results indicate well-smoothed trajectories with good comfort and safety attributes.

6. A Real-Time Detection Algorithm for Semi-structured Boundaries in the Park Based on 3D LiDAR

   Chao Ban, Ruijuan Chi, Tong Su, Yueqi Ma, Xiulian Huang

   Abstract

   Efficient and accurate real-time detection of semi-structured driving boundaries within the park is required by unmanned ground vehicles (UGVs) in order to facilitate vehicle positioning and path planning. These boundaries have a regular shape, but do not have obvious line and surface features, such as low shrubs. It is crucial for ensuring optimal navigation for UGVs. We have developed a new method for detecting semi-structured boundaries based on 3-dimensional Light Detection and Ranging (3D LiDAR), which involves three key steps. The first step involved pre-processing the 3D raw laser point cloud, which included three tasks: coordinate correction, segmenting the region of interest (ROI), and ground point cloud segmentation to achieve point cloud down-sampling. Next, the non-ground points in ROI were projected onto the ground, and boundary feature points that affect the driving of UGV were extracted. To eliminate incorrect boundary points, a two-step method consisting of the Random Sample Consensus (RANSAC) algorithm and inter-frame boundary line angle elimination was used. In the third step, the data of Position and Orientation System (POS) were integrated to construct a global point cloud map of the semi-structured driving boundary. An experiment was conducted on a self-developed UGV-LiDAR platform to evaluate the performance of the proposed algorithm. It was indicated by the results of the study that the feature point detection rate was exceeded by 98.95%, and the average processing time per frame was around 45.7 ms when there are more shrubs, meeting the real-time demands. This research provided valuable insights into perception technology for security operations of UGVs in parks.

7. Streaming Object Detection on Fisheye Cameras for Automatic Parking

   Yixiong Yan, Liangzhu Cheng, Yongxu Li, Xinjuan Tuo, Baoqiang Huang, Yakun Zhu

   Abstract

   Fisheye cameras are widely employed in automatic parking, and the video stream object detection (VSOD) of the fisheye camera is a fundamental perception function to ensure the safe operation of vehicles. In past research, the difference between the output of the deep learning model and the actual situation at the current moment due to the existence of delay of the perception system is generally ignored. Nevertheless, the environment will inevitably change within the delay, which may cause a potential safety hazard. This paper proposes a real-time detection framework with a dual-flow perception module (dynamic and static flows) that can predict the future and alleviate the time-lag problem. Meanwhile, we use a new scheme to evaluate latency and accuracy. The standard bbox is unsuitable for the object in fisheye camera images due to the solid radial distortion of the fisheye camera. The primary detection objects of parking perception are vehicles and pedestrians, so we adopt the rotated bbox and propose a new periodic angle loss function to regress the angle of the box, which is the simple and accurate representation method of objects. The instance segmentation ground truth is used to supervise the training. Experiments demonstrate the effectiveness of our approach. Code is released at: https://gitee.com/hiyanyx/fisheye-streaming-perception.

8. Research on Real Vehicle Testing of Path Planning for Autonomous Driving Tractors Under Ice and Snow Conditions

   Huaiguang Guan, Peng Guo, Jiansen Yang, Haipeng Xia, Dengquan Zhang

   ABSTRACT

   In order to test the safety of self-driving tractor-trailer vehicles when encountering snow and ice on low-grade roads, a convex corridor-based spatial path planning method adapted to snow and ice conditions is utilized. The method constructs a secure convex corridor in ice and snow regions by mapping obstacles and initial risks onto the Frenét coordinate system. It abstracts information such as safety, boundaries, and lateral control constraints, using the vehicle’s current position as the starting point. Spatial path optimization is then employed to generate a suitable and continuous speed trajectory. In light of the unique features of the ice and snow environment, a practical vehicle test was conducted on the Heihe test track. This test included four typical scenarios for automatic driving path planning on ice and snow road surfaces: cone bucket obstacle avoidance, pedestrian crossing, pedestrian following, and manhole cover. The experimental data collected during the test were processed in the tractor-trailer autopilot system. The results demonstrate that the method can effectively solve the quadratic planning problem with high robustness and practicality. This provides a valuable reference for the application of path planning algorithms in automatic driving technology, particularly in snow and ice areas.

9. Curve Trajectory Tracking for Autonomous Vehicles Using Linear Time-Varying MPC

   Shuping Chen, Huiyan Chen, Zhiguo Zhao

   Abstract

   Most researches concerning the yaw stability and rollover prevention for autonomous vehicles are studied separately and decoupled the longitudinal and lateral vehicle dynamics control. However, the roll motion influences the yaw stability during high speed curve steering manoeuvres. With this in mind, a curve trajectory stable tracking controller for autonomous vehicles using linear time-varying model predictive control (LTV MPC) is proposed in this paper. The lateral dynamics control employs the LTV MPC to generate a sequence of optimal steering angles considering the constraints of control input, state output, yaw stability and roll stability together in the cost function, in which the prediction model utilizes an 8 degrees of freedom (DOF) vehicle model and the plant utilizes a 14-DOF vehicle model. The longitudinal control adopts PID control embedded in the MPC framework to update the speed at each optimization step and generate the total wheel torque for speed tracking. The trajectory tracking simulation results demonstrate that the vehicle tracks the reference trajectory and speed well with the proposed controller, which satisfies the constraints of control input, state output as well as the boundaries of yaw stability envelope, sideslip angle and roll angle, thereby reducing the risk of vehicle skid and rollover.

10. Innovation and Optimization of VMT Visual System Calibration Method

    Zhou Zhao, Li Jinwei

    Abstract

    VMT visual localization is a technology that employs cameras for non-contact positioning of workpieces. It is commonly used in high-precision automated sealing workstations for body positioning. The traditional approach to calibrating visual systems has been through online debugging. This method requires using production downtime, has short debugging time windows, long cycles, and incurs high trial-and-error costs. Moreover, the necessary coordinate transformation techniques in visual debugging are often proprietary and require expensive outsourcing due to foreign vendor confidentiality. To address these challenges, this paper proposes two optimizations for visual calibration methods: 1.a novel offline debugging approach to replace the conventional online debugging, reducing the debugging cycle and enhancing efficiency. 2. an innovative coordinate transformation strategy and implementation method, along with a one-click transformation feature implemented through Python code. Lastly, the effectiveness of the proposed methods is demonstrated through application on new vehicle models. This leads to significant cost savings in debugging expenses. The approach has been successfully promoted within the company.

11. Energy Saving on Spray Booth Exhaust Treatment in Existing Factory

    Shaoyang Zhou, Lin Shen, Yin Hou, Ji Wang, Weijia Mao, Simin Deng, Xiao Zhang

    Abstract

    The continuous updated environmental protection laws on the exhaust gas treatment of the automobile industry has put forward specific guidance and stricter emission requirements. Automobile company follows environmental protection laws requirement, the treatment of the exhaust is from solvent borne paint to water borne paint. Deep treatment needs to consume a lot of electric and natural gas for spray booth’s exhaust have large volume and low concentration. To explore the energy saving treatment of the exhaust and realize the efficient operation of the spray booth and exhaust treatment system can help the automobile industry achieve carbon dioxide emissions peak and carbon neutrality.

12. Research on the Stamping Process of a Special Roof Panel

    Yi Bengang, Shen Xianfu, Chen Chenghua, Chen Jingjie, Hu Di

    Abstract

    This paper investigates a special roof panel as the research object and conducts a simulation analysis to study issues such as crack and wrinkle, and quality defects of A-class surface during the forming process. Integrating the requirements of die structure, welding, and sealing, the paper proposes feasible optimization schemes for the product structure and stamping process. After through trial production and validation, the parts meet the quality standards, achieving a perfect combination of product design and manufacturing process.

13. Full Four-Door Universal Handing JIG and Products Based on MTP System

    Dezhi Zhao, Yong Wang

    Abstract

    In the process of multi-vehicle door collinear flexible production in the automobile companies, the handing tools are dedicated to ATC (automatic tool changer) and have complex structure, large area and complex line body. Taking the four-door flexible collinear production of Dongfeng Honda Automobile Co., LTD as an example, combined with the product design characteristics, MTP (Manufacturing and Technical Proposal) system was integrated and set by D&E (Designer & Engineering) to confirm the connection handing jig and sheet metal design features, and finally designed the four-door collinear flexible universal handing jig and sheet metal features of door products. The results show that the setting and application of universal MTP can effectively simplify the handing jig structure, eliminate the handing jig, reduce the occupied site, reduce the cost and the time of subsequent new model synchronous engineering and equipment debugging, especially for the early cost reduction of new plant construction.

14. Study for Commercial Vehicle Cab Occupant Protection Standard of GB 26512-2021 and Development Test

    Bin Yang, Qi guang Sheng, Hui-yun Zhang, Bin Lei, Jiaqi Li

    Abstract

    The national standard GB26512-2021 of “The protection of the occupants of the cab of commercial vehicles” brings challenges to automobile enterprises to meet certification requirements. This paper introduces the technical route of revisions of GB26512-2021, and analyzes the technical requirements of GB26512-2021 for automobile test institutions and commercial vehicle manufactures. This paper introduces the technical route of revisions of GB26512-2021, and analyzes common key issues of test method of this standard in the course of implementation. This paper introduces the technical route of revisions of GB26512-2021, and analyzes common key issues of test method of this standard in the course of implementation. Based on the basis of a large number of test dates and typical test cases with the domestic manufactures, the paper introduces the solution that the cab of commercial vehicles adopts high strength frame structure, then conducts a research and analysis on the specific structure of the typical test problem cases, then analyzes the common key issues of test method of this standard in the course of implementation. Typical test problem cases, then analyzes the methods that the manufacturers improve key collision components at a lower cost. This paper provides some This paper provides some reference suggestions for solutions on safety standard of structural strength of commercial vehicle cab in china, and development test of commercial vehicle cab.

15. A New Technology of Concentricity Measurement Based on CMM

    Yunfeng Dai, Guiping Zhu, Bin Cui, Yi Zhu

    Abstract

    This paper analyzes the deviation of concentricity measured by coordinate measuring machine (hereinafter referred to as CMM). We can select a better measurement system and sampling strategy, so as to improve the accuracy of concentricity measured by CMM to the greatest extent. It is the first time to put forward the concept of uncertainty in measuring concentricity with different CMM under specific conditions. This paper analyzes the advantages and disadvantages of various flexible measurement methods, also point out the calculation defects of common axis method and straightness method and the possibility of misjudgment. Finally, the author summarizes the experience of measuring concentricity with CMM.

16. Study on Vehicle Clearance Tolerance and Extreme Value Based on 6σ

    Jiulei Cao, Shaohua Li, Xiaoli Zhao, Qingsong Liu, Xiaoyan Xing

    Abstract

    Part deviation mean, the matching two parts, the mean deviation direction should be consistent, and the deviation is consistent; For mass-produced parts, there are fluctuations between individual and individual parts, The deviation mean trend is consistent (P’ small), It does not mean that the P value of the vehicle DTS target can be achieved. By matching the sample frame, from focusing on the qualified rate of single parts to the matching of associated parts, And to evaluate the matching of the door cover size tolerance, Improve the matching ability of overlays, Combined with the encrypted measurement data application, lock the problem points; Using optical scanning, the deviation value is evaluated by locating the fitting with the contour matching point. Through the fitting evaluation of the outer plate, the problems are intuitively found and rectified.

17. Discussion on the Evaluation Method of Maintenance Technology Based on Automotive Engineering Design

    Fuyin Yu, Hongtao Yang, Yaokai Zheng

    Abstract

    Maintainability, as one of the evaluation factors of automobile quality, will affect the purchase decision of users to a certain extent, which makes it more widely valued in product research and development. This paper puts forward the dimension, method and comprehensive evaluation model of automobile maintainability analysis from the perspective of technology, and evaluates the maintainability in the engineering design phase, providing a quantitative evaluation basis for the maintainability of automobile design.

18. Analysis of Abnormal Seatbelt Shoulder Force Under the Front Collision Condition

    Zhiping Shi, Jiawen Li, Hongji Chen

    Abstract

    The phenomenon of abnormal seat belt load limit force under front collision conditions was studied. Based on the seat belt restraint system model and FTA, the shoulder force in each stage under front collision conditions was analyzed, with a focus on analyzing various possible factors of abnormal shoulder force during the force limiting stage and FTA model for abnormal shoulder force was established. Based on a specific practical case, analyze the impact of each factor in the FTA model on abnormal shoulder force and eventually form a force limiting correlation table for the root cause. A comprehensive analysis of the abnormal seat belt shoulder force from the constraint system level is summarized, which can quickly and comprehensively identify the root cause for the abnormal shoulder force and has a certain reference significance in the industry.