Vehicle Control And Active Collision Avoidance Based On Decomposed Fuzzy PID

With the development of unmanned vehicle technology,the safety driving of the unmanned vehicle is gradually concerned by researchers.The unmanned vehicle control system is one of the main researches directions in the field of unmanned vehicles.The driving safety is the primary goal of unmanned vehicles,the research,and development of the active collision avoidance technology of unmanned vehicles are very necessary.According to the characteristics of active collision avoidance technology,this paper uses the Decomposition Fuzzy System(DFS)to adjust the PID parameters in real-time and then uses DFSPID to control the unmanned vehicle to achieve the purpose of collision avoidance during longitudinal driving.The dynamic model of the whole vehicle and the DFSPID controller is established,and four working conditions similar to the actual conditions are considered to verify the feasibility of the active collision avoidance system.(1)Because the dynamic model is needed to support the verification of the active collision avoidance system,the longitudinal dynamic model of the unmanned vehicle is built.The longitudinal power assembly of unmanned vehicles includes engine,hydraulic coupling,automatic transmission,and vehicle mass model.To verify the rationality of the dynamic model built in this paper,the whole vehicle model is also built-in CarSim and compared with the longitudinal dynamic model built-in Simulink to prove the integrity and accuracy of the dynamic model.(2)Unlike the traditional fuzzy system,DFS decomposes the fuzzy variables into N layers,except for the two boundary fuzzy sets.The traditional fuzzy variable of each layer consists of a corresponding fuzzy set and its complement,and each layer corresponds to a fuzzy set in the original fuzzy partition.Because there are too many rules in the expert experience-based fuzzy rule base built by DFS,the simple decomposition fuzzy system(SDFS)is adopted.For SDFS,compared with DFS,only the fuzzy sets from the same sort layer of fuzzy variables need to be considered,the number of fuzzy rules is relatively small.(3)For the active collision avoidance system of unmanned vehicles,the model of collision avoidance logic adopts a safe distance model.The state information of the front and rear vehicles is input into the controller,and the corresponding danger threshold isobtained according to the designed calculation method.The threshold is the amount indicating the danger degree of the current vehicle condition.The unmanned vehicle can identify the emergency danger condition and whether to trigger the vehicle independent braking measures through the threshold method.Once the vehicle's autonomous braking is triggered,the system will give the expected deceleration for braking,and the unmanned vehicle will brake with the desired acceleration as the target.The hierarchical control strategy of the control target design layer and torque output layer is adopted.In the target design layer,the expected longitudinal acceleration is compared with the actual value to get the corresponding error;in the torque output layer,the error calculates the torque required by the expected acceleration by decomposing the fuzzy PID controller Finally,the effectiveness of the proposed control method is verified by simulation.