Vehicle lateral motion estimation with its dynamic and kinematic models based interacting multiple model filter

초록

In this paper, we presents a comparative study of dynamic and kinematic vehicle models based on interacting multiple model(IMM) filter. It is known that the characteristics of kinematic and dynamic lateral motion models vary according to driving conditions. From the IMM filter, we can obtain the stochastically best blended state of the vehicle. Not only the performance of IMM filter but also the reliability of both kinematic and dynamic lateral motion models were evaluated according to the steering wheel angle, yaw rate and side slip angle. The performance of each model based filter and IMM filter were validated via experimental results with a test vehicle driven on a high speed circuit under various road and driving condition. We observe that the IMM estimation which reflects each model characteristic is robust against various driving conditions. Furthermore reliability of each model could contribute to the slip angle estimation which is very challenging to estimate accurately due to sensor cost and complexity.