Adaptive tracking of road vehicle engine speed based on acceleration measurements

The paper presents adaptive approach to a method of vehicle engine speed tracking based on acceleration measurements which are taken in the vehicle body part. Engine-induced vibrations, which are acquired using a vehicle measurement system, as strongly nonlinear can be modelled using multi-notch filter. It was stated the optimal parameter of the multi-notch filter exists; however, solution space function of the optimization problem is significantly nonlinear. For such conditions adaptive LMS algorithm, which includes multi-notch filter, tends to converge to local minimum points of the solution space. Cross-correlation estimation of multi-notch filter output and its delayed version as well as power estimation of engineinduced acceleration signal were used to tune online an adaptation constant of LMS algorithm and bandwidth parameter of multi-notch filter. Results obtained for stationary and nonstationary engine-induced vibrations justified high accuracy of the tracking algorithm.