Set-based multiple model estimation for uncertain systems with unknown-but-bounded uncertainties

Set-based multiple model estimation (SMME) is investigated in this paper for uncertain systems with unknown-but-bounded (UBB) uncertainties. First, a novel Gaussian-form compatibility function based on set information is introduced for model identification. Then, the estimation set is calculated by combining the model-weight update and set-membership estimation. Note that the normalized model weight of each candidate model is updated according to the current measurement, which can be employed to identify the uncertain parameters. By combining the P -radius method and the set of estimation error, the gain matrix can be designed such that the P -radius of the set of estimation error is ultimately bounded. Finally, a second-order vibration system with uncertain natural frequency is employed to illustrate the effectiveness of the proposed method.