Research On Control Strategy Of Assistance System For Misoperation Of Accelerator Pedal Based On Multi-Sensor Data Fusion

In recent years,with the rapid development of the automobile industry,the increasing road traffic accidents have brought enormous losses to people's lives and property.Road traffic safety has become a public safety issue of major concern in the world.According to incomplete statistics,traffic accidents caused by drivers accidentally stepping on the accelerator account for 12.6% of the total number of traffic accidents.Therefore,from the perspective of active safety,this thesis analyzes and studies control strategy of assistance system for misoperation of accelerator pedal base on multi-sensor data fusion.First of all,the main research content of the thesis is determined through reading the literature both domestic and aboard,from which the latest research status and deficiencies of assistance system for misoperation of accelerator pedal is investigated.By comparing the difference of pedal behavior in driver's normal acceleration and emergency braking,the parameters for judgement of misoperation are identified,which include the driver's accelerator pedal force,accelerator pedal acceleration and steering wheel grip strength.Since one of the main reason of driver's misoperation is the sudden emerge of obstacle,it is of great importance to detect the distance between the vehicle and the obstacle in real time.Then the general design scheme,which based on film pressure sensor,accelerometer and ultrasonic sensor,is fulfilled on the basis of central control unit with Raspberry Pi.And the design of circuit and the mechanical structure of the system was completed based on Proteus and Solidworks respectively.In addition,combining the pros and cons of the existing assistance system device,the machine-electric integrated device is designed without changing the original structure of the car,and the 895 DC motor is selected as the power source.Then,the multi-input and single-output(MISO)fuzzy inference system is established based on fuzzy theory.And the accelerator pedal force,accelerator pedal acceleration,the steering wheel grip strength information are fused with the designed fuzzy membership function and fuzzy rules,the probability of the driver's misoperation of accelerator pedal is determined as a result.Take the ranging information from the ultrasonic sensor into consideration,the control strategy of the actuator is established.To satisfy the demand of quick response and stable output of the assistance braking device,a DC motor model is setup in Simulink,based on which the PID control algorithm,the three closed-loop control system,namely "torque-current-speed",is designed.The correctness of the theoretical derivation and simulation model are verified by simulation,and the control parameters of each loop are set in the process of simulation.The joint simulation results show that the fuzzy inference system designed in this paper can obtain the reasonable probability of the driver's misoperation of accelerator pedal with regard to diverse situations.And the working requirements of the assistance brake device can be met with the actuator motor by selecting the appropriate PID control parameters.Finally,the assistance system for misoperation of accelerator pedal is built on the real vehicle,and the performance of fuzzy system as well as the actuator are verified by the parking test.As a conclusion,the multi-sensor fusion based assistance system for misoperation of accelerator pedal is of great significance to the improvement of safety of driving.