This paper presents a study on the state estimation of out-of-plane dynamics of motorcycles based on the Sharp 71 model. The Sharp 71 model is a linear time-variant system that describes the out-of-plane dynamics of a motorcycle. Comparisons with multi-body simulations and measurement data show that this relatively simple model is capable of principally representing the lateral dynamics of the motorcycle. Two relevant variables of out-of-plane dynamics are the roll angle and the tire lateral forces. The structure of the Sharp 71 model offers the possibility of estimating these two variables model-based with the aid of corresponding measured output variables. The input variable is the steering torque, which obviously cannot be measured with reasonable effort. Therefore, an unknown-input observer is used to estimate the states. This state estimator allows a systematic consideration of the unknown input variable. The unknown-input observer is designed for different sets of outputs and the corresponding effects on the results are considered. The sensors used include gyroscope, acceleration sensor and steering angle sensor. The longitudinal velocity as time-variant parameter considers the coupling of the out-of-plane model with longitudinal dynamics. The implementation is achieved with gainscheduling of the observer feedback. The implemented concept is evaluated regarding its performance and convergence. Simulation studies are used as well as an evaluation based on measurement data. The simulation test was carried out with the help of a multi-body simulation. A comparison with a purely IMU-based roll angle estimator is presented for the handling course test. The results are of great interest, since in modern driver assistance systems, knowledge of the current dynamic vehicle condition is essential. With the aid of turn rate and acceleration sensors, important parameters such as roll angle are already recorded. The use of comprehensive physical motorcycle models is a pursued approach to further detail the vehicle state estimates. In combination with modern control engineering methods, other important driving dynamics variables can be calculated, such as the lateral forces of the tyres in this case. The results based on the simple Sharp 71 model already yields stable estimates of the essential state variables. The results also give information about the necessary level of detail of the used physical model and allow a principal assessment of the observer convergence during high dynamic maneuvers.
|Journal||SAE Technical Papers|
|Publication status||Published - 2020|
|Event||SAE 25th Small Engine Technology Conference and Exposition: Small Powertrains - Innovating for Their Future Role, SETC 2019 - Hiroshima, Japan|
Duration: 19 Nov 2019 → 21 Nov 2019