A fusion estimation of tire vertical forces using model-based tire state estimators for a dual-sensor intelligent tire

Abstract

Estimating tire vertical forces is essential to vehicle state estimation and stability control. Intelligent tires can be used to estimate tire vertical forces, but functional safety and extensive tests are issues to consider during intelligent tire development. This paper proposes a fusion estimation approach using model-based tire state estimators (TSEs) to estimate the tire vertical forces of a dual-sensor intelligent tire, which can output the circumferential strain, radial, and circumferential acceleration signals with a strain sensor and an accelerometer mounted at different positions on the inner liner. The mutual conversion between strain and acceleration signals is indicated in this paper for the first time; therefore, the internal relationship between different signals is revealed. Each measurement signal of the two sensors corresponds to a TSE composed of a signal processing algorithm, a mathematical model, and a Kalman filter. The mathematical model is proposed in this paper based on the flexible ring tire model (FRTM). The final estimated value of the tire vertical force is obtained by weighting and summing the outputs of the three TSEs. The weighting factors are determined using the genetic algorithm to study the fusion estimation effect. An integrated CarSim model is built in this paper to validate the estimation performance under various driving conditions, including driving straight at a constant speed, driving on an S-shaped road, and performing a double lane change at a high vehicle speed. For all driving conditions, the mean error rates of the fusion estimation are less than 2%. The model-based tire state estimators can avoid the extensive tests needed in the data-based methods. Furthermore, the fusion of the outputs of three TSEs can further improve the estimation performance compared with the situation when a single TSE is used. Therefore, the studies in this paper have guiding significance for intelligent tire development.