A Fusion Algorithm for Estimating Time-Independent/-Dependent Parameters and States

Abstract

Vehicle parameters are essential for dynamic analysis and control systems. One problem of the current estimation algorithm for vehicles’ parameters is that: real-time estimation methods only identify parts of vehicle parameters, whereas other parameters such as suspension damping coefficients and suspension and tire stiffnesses are assumed to be known in advance by means of an inertial parameter measurement device (IPMD). In this study, a fusion algorithm is proposed for identifying comprehensive vehicle parameters without the help of an IPMD, and vehicle parameters are divided into time-independent parameters (TIPs) and time-dependent parameters (TDPs) based on whether they change over time. TIPs are identified by a hybrid-mass state-variable (HMSV). A dual unscented Kalman filter (DUKF) is applied to update both TDPs and online states. The experiment is conducted on a real two-axle vehicle and the test data are used to estimate both TIPs and TDPs to validate the accuracy of the proposed algorithm. Numerical simulations are performed to further investigate the algorithm’s performance in terms of sprung mass variation, model error because of linearization and various road conditions. The results from both the experiment and simulation show that the proposed algorithm can estimate TIPs as well as update TDPs and online states with high accuracy and quick convergence, and no requirement of road information.

Precise Tracking Control for Articulating Crane: Prescribed Performance, Adaptation, and Fuzzy Optimality by Nash Game

“…This leads to a classical problem, that is, how to mathematically describe uncertainties. Hitherto, the probability theory is the most used method to describe uncertainties [22], [23]; however, it may not be well suited to describing the vast majority of uncertainty as it demands a large amount of experimental data [24]. In contrast, the fuzzy set theory interprets the uncertainties by degree of occurrence, which can dramatically reduce the need for data amount through expert opinions [25], [26].…”