Tobit Recursive Filtering for Nonlinear Networked Systems with Measurement Censoring and Constrained Bit Rate

The Tobit recursive filtering (TRF) issue is investigated for nonlinear networked systems with measurement censoring (MC) and constrained bit rate. The MC considered is characterized by the Tobit Type I measurement model. The constrained bit rate model is used to describe the bandwidth allocation mechanism of the digital channel. The encoding-decoding strategy is elaborately adopted to ensure the reliability of data transmission under the constrained bit rate. A Tobit recursive filter is designed such that, for both the possible MC and constrained bit rate, an upper bound (UB) of the filtering error covariance can be found by solving the recursive matrix equations. Subsequently, the expression form of the filter parameter matrix is provided by optimizing the trace of the obtained UB. Additionally, the monotonicity of the minimized UB in relation to the bit rate is analyzed. Finally, the validity of the TRF strategy is demonstrated through a simulation example.