Transfer Case Clutch Torque Estimation Using an Extended Kalman Filter With Unknown Input

Vehicle wheel traction forces play an important role in vehicle performance, especially for four-wheel-drive vehicles, where a transfer case clutch is used to distribute torque between front and rear wheels. Due to lack of production ready low-cost torque sensor, the transfer case clutch torque is not measured and needs to be estimated accurately to optimize vehicle traction performance. This article proposes to estimate the transfer case clutch torque by forming an estimation problem with unknown system input(s). Specifically, a unified clutch torque estimation model is developed for different clutch conditions, where overtaken-clutch case is treated as a special case of slip-clutch condition with slip-speed equals to zero. Note that under overtaken condition, the associated system model with only one known input is different from that under slip condition with two known inputs. A real-time implementable recursive solution for unknown input(s) is obtained by utilizing an unknown input observer based on the extended Kalman filter (EKF-UIO) algorithm. Comparing with the existing direct estimation method and experimental measured data, it is found that the proposed EKF-UIO algorithm is able to reduce both absolute mean square error and relative-mean-square error significantly with respect to the measured clutch torque under both slip and overtaken conditions. In summary, the EKF-UIO estimation algorithm based on the unified clutch torque model is able to estimate clutch torque accurately.