Study On The Driver Model And Objective Evaluation Of Tire Performence For Vehicle Handling

CAE is an effective development tool that can reduce the iterative process of thedevelopment process. Before the prototype vehicle is available, CAE can reduce costs ofdevelopment and short the development period. Study on the driver model and objectiveevaluation of tire performance for vehicle handling is important for reducing tire matchingcosts and time.As the only component to translate the forces and moments between the tire and theroad, it affects dynamic performance of the vehicle. Some performance is difficult to make acomplete evaluation only by its own characteristics. Because tire is the important part ofvehicle coupling with vehicle, especially the vehicle’s steering system, suspension system,braking system and powertrain, the performance of tire can be effectively evaluated with thevehicle.In this paper, tire performance objective assessment is the key point. The work includethe tire model and vehicle model, internal model of MPC, the driver’s classification,neuromuscular model, closed-loop feedback control, objective evaluation of vehicle handling.The most work is summarized as follows.1. The model of vehicle is built for the simulation of the vehicle handling. The accuratevehicle model is the foundation of the vehicle handling simulation. In this paper, theobjective evaluation of tire performance is studied for vehicle handling ability. The highprecision of tire model is needed. Take into account both tire model accuracy andcomputational cost, the semi-empirical UniTire model is selected for this work. The mechanical characteristics of4types of tire are tested. The parameters of UniTire model areidentified by experimental data. The K&C characteristic of suspension and steering systemare measured on prototype by K&C rig. The inertia and mass about the vehicle are measured.The vehicle dynamics model is established using CarSim software. The test is carried toverify the accuracy of the model. The comparison between the experimental results andsimulation show that the accuracy of the model can meet the needs of the follow-up study.2. The personalized driver model is built. Driver model is an important component indriver—vehicle closed loop system. The driver’s experience and personality are different. Thesame vehicle shows different performance with different drivers.The optimal driver modelcannot express the personality of the driver. The drivers are divided into novice drivers andskilled drivers based on the experience of driver. The awareness of novice drivers about thevehicle lateral dynamics is limited in the linear range. Skill driver’s experience can beextended to non-linear range. The relax driver with relax muscle and focused attention canwell control car. Nervous driver with inflexible muscle and sensitive reaction can’t drive carwell. MPC controller and close-loop feedback control combined driver model neuromuscularmodel can simulate the driver’s steering behavior. This driver model consists of MPCcontroller, the closed-loop feedback control and neuromuscular models. MPC controllergenerates control commands to characterize the manoever of driver by the experience ofdriver. Closed-loop feedback control command to characterize the correction of driver bytolerance of current position and attitude. The steering angle consisted of the two parts isapplied to the steering wheel through the arm neuromuscular models. The reference model ofMPC controller using a linear single-track model express the experience of novice driver. TheMPC controller with a set of single-track model extended to a non-linear dynamics rangeexpress skilled driver. The genetic algorithms are applied to identify model parameters. Thecornering stiffness of its front and rear axle are identified. The MAP is established by the dataof cornering stiffness, the speed and steering wheel angle as variables, cornering stiffness ofaxle as dependent variable. The internal model of MPC is determined by the MAP based onthe vehicle status. The driver’s states are distinguished by the inertia, damping and stiffness of the driver arm neuromuscular model. DLC and slalom simulation tests from low to highspeed show that the proposed driver model can follow the desired path. Contrast the novicedriver and skilled driver with120km/h DLC simulation shows that the MPC control andfeedback compensation combined with neuromuscular model can realize the driverclassification. Contrast the driver model proposed in this paper to MPC control driver modelwith neuromuscular model with100km/h DLC, the proposed driver model can better trackthe desired path and reflect the driver’s the correction with the vehicle position and headingangle error. The proposed driver model can reflect the driver’s personality than optimal drivermodel.3. The vehicle dynamics model and driver model are applied to constitute driver-vehicle closed loop system. The system is used to objectively evaluate vehicle handlingability. In this paper, the vehicle model of vehicle equipped with4types of tire and drivermodel is used to simulate double lane change and slalom test. Simulation results show thatthe parameters about vehicle state with different tire are not consistent. In order to evaluatethe tires’ performance, the comprehensive evaluation index of vehicle handling is adopt toevaluate closed-loop system. The simulation results show Atire is the worst for the vehicle inthe4type tires. C tire is just better than A tire. The performance of B tire close to D tire.There is desired vehicle response, after he performs a manipulation. If the difference betweenvehicle response and desired is small, then predictability of vehicle handling ability isconsidered to be good. The changed trend of vehicle handling ability evaluation from lowspeed to high speed is used to express the predictability of the vehicle handling stability. Ifthe trend is smoothly, then the driver can well master the vehicle handling ability and thepredictability of vehicle handling ability is good. The trend of objective evaluation of4typesof tire change with the speed show the predictability of performance of D tire is the best inthe4types of tire.4. To verify the objective evaluation of closed-loop system，the handling of vehicleequipped with4kinds of tire is subjectively evaluated by2drivers. The most evaluation oftwo drivers about4kinds is consistent. Considering the discreteness of the drivers, the test results of subjective evaluation are acceptable. Experimental results illustrate the goodconsistency between the proposed objective evaluation and subjective evaluation.This article has the following innovation points.The MPC controller and closed-loop feedback control combined with neuromuscularmodel are used to model driver. The single-track of vehicle is used to express the driver’sexperience. The linear model is used to express the limited experience of novice drivers. Thenonlinear model is used to express the experience of skilled driver. Simulation of DLC andslalom shows that different parameters of neuromuscular model reflect the driver’smanipulation of state. Simulation of novice driver and skilled driver show that the differentscope of reference model of the MPC controller can reflect the driver’s experience. Comparetwo driver models shows that the proposed driver model can better tracking expected path,and can reflect the driver’s correction with the vehicle position and heading angle error.The handling ability of vehicle equipped with4types of tire are objective evaluatedfrom low speed to high speed. The performance of4types of tires are objective evaluated byrelax and tense driver. The consistency of tire handling is evaluated by the trend of tirehandling ability with the change of the speed. Compare the results of simulation and testshows the objective evaluation and subjective evaluation are agreed.