Dynamic quantization for track fusion under communication constraints

A new quantization approach is proposed for track fusion in fusion systems under communication constraints. The quantization algorithm used in practice for track fusion is a static nearest neighbor approach which selects the closest vector and the covariance in a table to the current track information. The quantization algorithm proposed here involves posing the quantization problem in an optimization framework and solving it by also including the predicted future values of the track into the picture. Since the approach considers the inherent dynamic characteristics of the tracks, the resulting methodology is called as dynamic quantization. The early simulation results show that the dynamic quantization is much more advantageous compared to static quantization even under very low bit rates.