Research On Vehicle State Estimation Algorithm Based On Fuzzy Model

With the increase of vehicle ownership,people have been widely concerned active safety and handling stability of vehicles.Active safety control system can effectively improve vehicle handling stability and avoid traffic accidents.These active control systems can effectively improve the safety of vehicle driving on the premise of accurate vehicle driving state information,that is,vehicle longitudinal speed,lateral speed and yaw speed.Most of these information is obtained by the sensors installed on the vehicle,but the lateral velocity of the vehicle is difficult to accurately obtain by the on-board sensors,which makes it impossible to measure.In nonlinear system theory,an important solution is to model vehicle nonlinear system as Takagi-Sugeno(TS)fuzzy model.However,the traditional vehicle TS fuzzy model is composed of local linear subsystems.If the vehicle lateral velocity is unmeasured,the vehicle TS fuzzy model antecedents are unmeasured,and the "mismatch" error term will appear in the process of observer design,which makes the design of the observer more challenging.1.In this paper,a new fuzzy model modeling method,namely vehicle N-TS fuzzy model,is proposed.Firstly,the nonlinear terms of the vehicle system are divided into measurable and unmeasurable terms,and then the measurable nonlinear terms are divided into sector nonlinear methods,and all the unmeasurable terms of the system are regarded as the local nonlinear sub-terms of the model.This avoids the appearance of unmeasurable premise variables and effectively solves the difficulty in the research of vehicle TS fuzzy system observer with unmeasurable antecedent variables.Because the model only adopts the sector nonlinear method to build the vehicle system fuzzy model for the measurable premise variables,the vehicle N-TS fuzzy model has fewer fuzzy rules and less computation.2.The vehicle state estimation algorithm based on Luenberger fuzzy observer is studied.Firstly,the vehicle N-TS fuzzy model and Luenberger fuzzy observer without unknown inputs are established.Then,the design conditions of the fuzzy observer are given according to Lyapunov stability theory to estimate the vehicle system state.3.The vehicle state estimation algorithm based on unknown input fuzzy observer is studied.Firstly,the vehicle N-TS fuzzy model with unknown inputs was established,and based on the model,the vehicle fuzzy system observer with unknown inputs was designed.The convex condition of the fuzzy observer was given,and the simultaneous estimation of vehicle system state and unknown inputs was realized.