Comparison of quantized state estimators with different transmitted information forms

Bandwidth limitation is an unavoidable constraint when data is transmitted from local sensor to the estimation center in networked systems. As a result, quantization strategy is often used to deal with this constraint during the design of networked state estimators. In this paper, we compare the performance of three quantized estimators with different transmitted data forms, such as the original measurement, the innovation and the local estimation. Firstly, adaptive bit quantization is introduced to deal with the bandwidth limitation constraint. Secondly, three quantized state estimators are introduced. Actually, they adopt the same quantizing strategy. Intervals and common variance upper approximation method are also used. Thirdly, we compare estimation accuracies of the three quantized estimators by using their estimation error co-variances. Finally, a simple simulation is demonstrated to validate the conclusion in our comparison. The results show that these three quantized filters have very similar estimation accuracy.