Comparison Study of State Vector Estimation Methods for Moving Objects

The paper considers the comparative study of different methods for state vector estimation of moving objects in conditions when data from sensors received with a low frequency, and the measurements themselves are very noisy. The Kalman filter is used for state vector estimation. Due to the low frequency of updating the measurement vector, it is proposed to increase the frequency of predicting the state of a moving object between these updates. The problem of identifying a moving object at noisy measurements is solved using three different methods: the nearest neighbor method, the probabilistic method of data association, and fuzzy logic. A comparative of the simulation results of the developed algorithm using the three considered methods is presented. When using fuzzy logic, significantly fewer false measurements are selected than with the other two methods, since the measurement of a moving object is selected based on several parameters of the object, but not only based on one parameter.