Performance bounds for the Kalman filter with compressed quantized measurements

This paper analyzes the performance of a Kalman filter which receives compressed quantized measurements. Recursive relationships are obtained for the propagation of upper and lower bounds to the expected error covariance of the filter estimate. The bounds represent the actual error covariance of possible alternatives to the optimal filter and can serve as analytical tools for comparing the performance of the filter, under different compression schemes, prior to the actual processing of the real-time data.