Estimation Approach Of Road Friction Coefficient For Electric Vehicle Active Safety Systems

With the funding of Shenyang science project (F12-277-1-11) of “The KeyTechnology of Doubly-Fed Motor, AWD, Differential, Handling and Stability Control ofElectric Vehicle”, this thesis adopts front-wheel drive electric vehicle as the researchobject, studies road friction coefficient estimation methods under various conditions.Firstly, a electric vehicle model with seven degrees of freedom is built underMatlab/Simulink platform which includes tire model, load transfer model. Compared withstandard vehicle model which is in vehicle dynamic simulation software Carsim underdifferent operating conditions, the electric vehicle model is verified valid. Underacceleration, braking and steering conditions, the influence to vehicle states with differentroad conditions is analyzed.Secondly, according to different operating conditions, road friction coefficientestimation based on vehicle lateral dynamic and longitudinal dynamic methods areproposed.Road friction coefficient estimation based on lateral dynamic approach applies tosteering condition includes self-aligning moment based and nonlinear observer basedestimation methods. Self-aligning moment based method which utilizes the self-aligningmoment curve characteristic to estimates road coefficient. Nonlinear observer basedmethod constructed a nonlinear observer which takes advantage of vehicle lateral dynamicmodel, the error between the measurements of vehicle sensors and the outputs of vehiclemodel is regarded as observer’s feedback terms, and the observer gain is confirmed.Road friction coefficient estimation based on longitudinal dynamic approach appliesto straight driving condition contains linear range and saturation range estimation methods.In linear range, based on longitudinal dynamic model, by the relationship betweennormalized force and slip ratio, estimation the slope of curve is implemented. In saturation range, according to traction force and road friction coefficient relationship, estimates roadfriction coefficient using least square method.Lastly, according to the accelerating and braking operation during turning condition,the nonlinear observer estimation method is improved by applying the gain reductionfunction. And the road friction coefficient estimation methods which based on longitudinaland lateral dynamic models have been integrated. When the vehicle is under differentoperating states, it can switch different estimation methods. Under straight and steeringconditions, the simulation results show the integrated estimation method has a goodperformance.