A High Effective Fuzzy Synthetic Evaluation Multi-model Estimation

Abstract

In view of the questions that the algorithm flow of variable structure multi-model method (VSMM) is too complex and the tracking performance is inefficient and therefore it is so difficult to apply VSMM into installing equipment. The paper presents a high-performance variable structure multi-model method basing on multi-factor fuzzy synthetic evaluation (HEFS_VSMM). Under the guidance of variable structure method, HEFS_VSMM uses the technique of multi-factor fuzzy synthetic evaluation in the strategy of model set adaptive to select the appropriate model set in real time and reduce the computation complexity of the model evaluation, firstly. Secondly, select the model set center according to the evaluation results of each model and set the property value for current model set. Thirdly, choose different processes basing on the current model set property value to simplify the logical complexity of the algorithm. At last, the algorithm gets the total estimation by the theories of optimal information fusion on the above-mentioned processing results. The results of simulation show that, compared with the FSMM and EMA, the mean of estimation error belonging to position, velocity and acceleration in the HEFS_VSMM is improved from -0.029 (m), -0.350 (m/s), -10.051(m/s2) to -0.023 (m), 0.052 (m/s), -5.531 (m/s2). The algorithm cycle is reduced from 0.0051(s) to 0.0025 (s).

Keywords

• Maneuvering target tracking
• Variable structure multiple-model algorithm
• Fuzzy inference
• Multi-factor evaluation.