Research On The Design And Test Of Automatic Parking Strategy For Intelligent Vehicle

As a key technology in intelligent vehicle,automatic parking can help drivers to park and improve people’s travel experience,which is an important research direction at present.Reasonable automatic parking strategy can improve the safety and comfort of the parking process,and improve the parking efficiency.However,with the increase of car ownership,the complexity of the parking environment and the difficulty of parking are also increasing,and high-quality parking path planning and accurate path following control are still the technical difficulties of automatic parking strategy.While the automatic parking strategy is developing,the test method is not mature.Scientific and efficient test methods will become an important thrust for the development of automatic parking technology.This paper,which is relies on the 13 th Five-Year National Key R&D Program Project "Research on Hardware-in-the-Loop Test Platform Construction and Simulation Test Technology for Automated Vehicles(2018YFB0105103)" and the National Natural Science Foundation of China Project "Simulation Test and Evaluation Method of Automated Vehicles in the Snow and Ice Environment(U22A20247)",focuses on the design and test of the automatic parking strategy.The automatic parking strategy includes path planning strategy and tracking control strategy.The path planning strategy searches for a high-quality path in the parking scenario by means of front-end path search and back-end path optimization.And then constructs lateral and longitudinal controllers to achieve parking path tracking.Using the scenario-based testing method,firstly,the elements and the range of values of the automatic parking strategy test scenarios are clarified,and then introduce the scenario attention as a guide to generate test scenarios that meets the design requirements based on the combinatorial testing.Finally,the overall verification of the automatic parking strategy is carried out,and the improvement effect of introducing scenario attention on the combinatorial testing scenario is verified.The specific research content of this paper are as follows:Automatic parking strategy design.The automatic parking strategy mainly considers the parking path planning strategy and tracking control strategy after the known environmental information.The front-end initial path was obtained by combining hybrid A* and multicurvature Reeds-Sheep curve set.Then,according to the path length,the model predictive control method or the discrete point reference line smoothing algorithm are selected to perform back-end optimization on the initial path to achieve high-quality parking path planning.In the tracking control strategy,the T-shaped and S-shaped speed curves were used to design the speed for the parking path to ensure the continuity of speed and acceleration.Thereafter,the lateral and longitudinal controllers with good tracking control effect are constructed,and the vehicle longitudinal speed control is realized by the combination of dynamics feedforward and feedback;the vehicle lateral path tracking control is realized by the Pure Pursuit algorithm.Automatic parking strategy test scenario generation.Firstly,the scenario elements are extracted for the characteristics of the automatic parking strategy,and the values of the scenario elements are clarified with reference to the standard regulations and realistic conditions;Then,the combinatorial testing was used to generate test scenarios,Analyzing the shortcomings of the traditional continuous particle swarm algorithm in solving the test suite generation problem,a discrete particle swarm algorithm is proposed to generate a smaller number of test suite under the premise of satisfying the given test coverage requirements;introducing the scenario attention to quantitatively describe the importance of testers to different scenarios,and using it as a guide to transform the test suite into automatic parking strategy test scenarios that better meet the design requirements of testers.Testing and verification of automatic parking strategy based on scenario testing.Firstly,the overall function of the automatic parking strategy was verified on the pure simulation test platform.After that,the automatic parking strategy is traversed in the test scenarios generated by the combinatorial testing based on the scenario attention guidance and the test scenarios generated by the combinatorial testing without considering the scenario attention,recording the front-end path search time,number of vehicle direction changes and tracking control error of the automatic parking strategies in the two scenarios,evaluate the difficulty of the scenarios and compare the test values of the two scenarios;Finally,in order to test the performance of the automatic parking strategy under close to the actual working conditions,a prototype controller-in-the-loop testbed is built to test and validate the automatic parking strategy.The results show that the proposed automatic parking strategy can effectively complete the parking task in more complex parking scenarios.And the scenarios generated by introducing the scene attention as a guide in the combinatorial testing method are easier to find the defects and shortcomings of the algorithm,which has higher test value.