Research On Safety Of The Intended Functionality Of AEB Based On System Theoretic Process Analysis Method

Safety has always been a core element of the sustainable and healthy development of the automotive industry.With the increasing emphasis on safety issues,safety of intended functionality（SOTIF）has emerged as the times require,that is,there is no unreasonable risk due to the inadequacy of the intended function or its implemented function or the harm caused by reasonably foreseeable human misuse.At present,the research on SOTIF is still at the theoretical level,and it is difficult to realize its landing research.In addition,with the continuous advancement of automotive technology,autonomous vehicles rely on technologies such as big data,artificial intelligence,computer vision,high-precision maps,high-precision positioning,and collaborative computing to gradually improve the level of vehicle intelligence.As a common active safety technology in autonomous vehicles,automatic emergency braking（AEB）is also very important for the safety of vehicles and occupants.Therefore,this article takes the AEB system as the carrier,analyzes and studies the various processes of expected functional safety,carries out the AEB design,adopts the STPA（system theoretic process analysis）method to analyze the safety of the AEB system SOTIF,and finds out the relevant information through batch simulation.Verify the security boundary of the scenario and the insecure scenario.The parameters of the AEB algorithm are adjusted by the genetic algorithm,and some unsafe scenarios corresponding to the AEB system are transformed into safe scenarios,the safety boundary is enlarged,the known safety area is enlarged,and the SOTIF of the AEB system is improved.First of all,relying on the AEB system as a carrier to design an analysis framework for expected functional safety,describe functions and system specifications,carry out hazard identification and risk assessment,and identify and evaluate related trigger events.According to the algorithm of AEB system,the verification methods and strategies of SOTIF are formulated,the verification scenarios are selected,and the measures to improve SOTIF are analyzed and studied.Then,according to the architecture of AEB system,the overall scheme of AEB is designed.According to the forward target information obtained by the millimeter-wave radar sensor,the most dangerous target selection algorithm is designed,and the most dangerous target is selected from many dangerous targets,considering the two scenarios of straight road and curved road.Besides,the hierarchical early warning/braking strategy of the AEB system was designed,the relevant parameters such as braking intensity and TTC threshold were determined,and related models were built,and PID was used to follow the speed of the main vehicle.The rationality of the AEB system in this paper is verified.Then,the safety analysis of AEB system SOTIF was carried out using STPA.According to the AEB system,the system-level hazards are defined,hazard identification and risk assessment are carried out,the security constraints of the system are clarified,and the hierarchical control structure of AEB is established.The hierarchical structure is modeled in more detail for the urban scene and the highway scene.According to the established hierarchical control structure in the two scenarios,the control behaviors that may lead to danger are identified,and the unsafe control action is identified,refine the safety constraints,and then identify the causal scenarios in the two scenarios,and select and evaluate the triggering events of the highway scenario.Finally,according to the selected verification scenarios,the test method of batch cosimulation of Prescan and Simulink is used to find out the safety boundaries of the straight and some curved scenarios in the designed AEB system,and to find out the relevant AEB false triggering scenarios.By improving the AEB algorithm,that is,the key parameters in the AEB algorithm are adjusted and optimized through the genetic algorithm,and finally the safety boundary of some curve scenes is expanded,the occurrence of AEB false triggering scenarios is avoided,and the unsafe area is transformed into a known safety area,so as to achieve the purpose of improving the intended functional safety of the AEB system.