A novel underwater-target tracking algorithm with experimental study

In view of practical applications of underwater tracking of moving-targets, the rate mea-sure obtained with sonar is a relative radial rate. It cannot be used directly, and the sensor data are color contaminated. A new moving target tracking algorithm with an extended measurement vector (i.e. relative radial rate) and an extended state vector is proposed in this paper to improve steady-state estimation accuracy and convergence rate. A new experimental scheme of tracking multi-target in a water-tank environment is presented. Comparison of performance between the proposed method and the measurement conditioned (MC) approach is given. The results show that there is significant improvement in the tracking capability over MC method.