Estimating Motorcycle Telescopic Fork Suspension Travel Speed with Four-Degree-of-Freedom Full-Vehicle Kalman Filter

This paper concerns the estimation of front telescopic fork suspension elongation speed through the use of Kalman-filtering techniques. A full-motorcycle model in the state-space domain is developed and subsequently used in the filter along with synthetic input data simulating two accelerometer measurements. In addition, the force of semi-active suspension is considered as an input, from which the value is estimated on the basis of a look-up table and the estimated elongation speed. The performance of the full-motorcycle filter is compared to that of a filter built considering the monocorner model, indicating superiority in performance. The ratio of the mean squared error of the suspension elongation speed to the mean square of the elongation speed originating from the non-linear model is used as a performance metric. For the proposed estimator, it is 6.54% with respect to the best class of road profile (A) and 7.07% for the worst (H). This is in contrast to the monocorner filter, displaying values of 57.46% and 94.47% for the best and worst road classes, respectively. The influence of system pitch dynamics is evidenced to have a marginal influence on the accuracy of speed estimation. However, it is the use of a larger set of states that adds the notable advantage of employing such a solution.