Rack force estimation from standstill to high speeds by hybrid model design and blending

Abstract

One of the main challenges of steer-by-wire (SbW) systems is force feedback, which requires either rack force measurement or estimation. In this work, two models are blended to estimate the rack force within the whole velocity range from standstill to high speeds. At standstill, a stationary steering model is presented and adjusted via Gaussian process regression (GPR) for a more accurate estimation. The lateral vehicle dynamics are used to estimate the rack force via an extended Kalman filter (EKF) while the vehicle is moving at high speeds. Both models are combined by two fitting approaches to allow for estimation at low speeds, at which each one shows unsatisfactory results. The resulting blended models are validated with measurements from a test vehicle.