Comprehensive Control Strategy For Distributed Driving Electric Vehicle Based On Double Layer Architecture

Distributed driving electric vehicle(EV) is an important direction of the futuredevelopment of electric vehicles. A comprehensive control strategy based on double-layerarchitecture is proposed for distributed driving EV. The upper is dynamic control strategylayer and the lower is energy optimization control strategy layer. The decoupling controlbetween the dynamic and energy optimization is realized through the control allocationlinear mappings between the upper virtual control and the lower actual control. The maincontent of this dissertation includes the following elements:Firstly, a simulation system for distributed driving EV is constructed according to thecharacteristics of dynamics control for distributed driving electric vehicle, so as to provideclosed-loop simulation environment for the conduction of comprehensive control strategy.The simulation system consists of vehicle model with7degrees of freedom, driver model,and tire model, etc. The precision and accuracy of the simulation system has been verified.Therefore, the simulation system is able to satisfy the need of this study.Secondly, the dynamics control strategies of the upper layer are conducted andtracking control strategy of the motion state of the vehicle routine’s tracking controlstrategy is focused on. Based on the sliding mode control theory, a tracking controlalgorithm of longitudinal velocity, lateral velocity and yaw velocity is designed, which isalso combined with analysis of drivers’ intents. The linear mapping relationship betweenthe virtual control generalized from the upper sliding mode control and the actual controlinput of lower layer according to the theory of control allocation. The parameters needed inthe sliding mode controller are not easy to obtain, so the algorithm on the improvedadaptive control is used to realize the estimation of the vehicle mass and moment of inertia.And simulation results show the effectiveness of the algorithm.Then, the energy optimization control strategies of the lower layer are studied. In thevehicle straight movement condition, the front and rear motors are taken as two integrationpairs. The efficiency of the motor is expressed by the piecewise linear functions. Takenboth the power consuming and the offset between the actual contol and the virtual controlas optimize issues, the energy efficiency optimization objective function is obtained on thebasis of the theory of nonlinear programmin. For the optimization objective function, the energy efficiency optimization algorithm in terms of KKT(Karush-Kuhn-Tucker)optimization conditions is presented and verified through the simulation,which canobviously reduce energy consuming.In the vehicle lateral movement condition, the efficiency of the motor is expressed bythe five order polynomial expressions. Taken both the power consuming and the offsetbetween the actual contol and the virtual control as optimize issues, the energy efficiencyoptimization objective function is obtained on account of the theory of nonlinearprogrammin.The adaptive update law is used to solve the objective function. The algorithmis verified through the simulation, which can obviously reduce energy consuming.Finally, the experimental verification is conducted. The hardware in the loopsimulation platform is designed and verifies the estimation results of vehicle mass andmoment of inertia based on the improved adaptive algorithm. Meanwhile, the trackingcontrol performance of sliding mode control on the longitudinal velocity, lateral velocityand yaw rate is verified through the real vehicle test on straight acceleration and cornering.