This paper presents a novel method to solve the problem of smoothing and prediction in real-time in marine radar tracking by means of a feedforward neural network, which is able to deal with any kind of target data. Varying steps are employed in the network training in addition to the momentum factor. Simulation results show that the predictive track error is rapidly convergent and the accuracy of stable tracking is good enough for marine radar data processing, especially for a ship with a speed below 2O nm/h. The network can also track the target in time which takes a sudden turn.
展开▼
