Multi-objective Collaborative Optimization Of The Differential Assisted Steering With Electric Motor

As an ideal power steering system, the Differential Assisted Steering System with electric motor,which can realize both active steering and assisted steering at the meantime, can not only combinesteering sense with steering portability perfectly but coordinate with steering safety and flexibility.Because the performances of the Differential Assisted Steering System were influenced andrelated with each other, it was essential to search for a Multi-objective Collaborative Optimizationmethod to deal with the optimization jobs. Based on the three degree of freedom handing stabilityvehicle model, the dynamitic model of the steering system was built and steering economy, steeringsense and steering flexibility were proposed and quantified. On the foundation of that, the parameters’influence on the performances was analyzed to choose the more influential ones to prepare for thenext work. After that, the theory about Response Surface Approximate Model based on LatinHyper-cube Design of Experiment was studied, and second order polynomial response surface modelsof every performance were built. By studying the Non-dominated Sorting Genetic Algorithm Ⅱ, theobjective function of the Differential Assisted Steering System and its constraints were built. Besidesthat, the Collaborative Optimization calculation structure, taken steering economy as main system andsteering sense, steering flexibility, the maximum torque angle of the sensor as subsystems, was built.The Multi-island Genetic Algorithm in the main system and the Sequential Quadratic Programmingalgorithm in the subsystems were adopted to collaboratively optimize the performances of the steeringsystem. Based on analyzing and comparing with Multi-objective Genetic Algorithm and CollaborativeOptimization, Non-dominated Sorting Genetic Algorithm Ⅱwas introduced into the main system ofthe Collaborative Optimization in order to translate the original single-objective optimization intomulti-objective one. The ideal Pareto optimization solution can be got after calculation.