Research on target tracking algorithm using improved current statistical model

摘要

Current statistical model actually is a modified Singer model, its mean value is the forecast value of current acceleration, the random maneuvering acceleration is still supposed to one order time correlative process in time axis. Because current statistical model can identify the maneuvering acceleration online and adjust the state noise covariance matrix, it is more close to reality compared to Singer model. However the maneuvering frequency usually is the experience value, which will result in random jump of the estimated acceleration and a big estimation error compared to actual situation if the target maneuvering is close to uniform motion. To solve this problem, a self-adaptive method for calculating maneuvering frequency was proposed in Current statistical model. The simulation result proved that the algorithm was valid.