In this paper, the mobile robot localization problem is investigated under the stochastic communication protocol (SCP). In the mobile robot localization system, the measurement data including the distance and the azimuth are received by multiple sensors equipped on the robot. In order to relieve the network burden caused by network congestion, the SCP is introduced to schedule the transmission of the measurement data received by multiple sensors. The aim of this paper is to find a solution to the robot localization problem by designing a time-varying filter for the mobile robot such that the filtering error dynamics satisfies the H∞ performance requirement over a finite horizon. First, a Markov chain is introduced to model the transmission of measurement data. Then, by utilizing the stochastic analysis technique and completing square approach, the gain matrices of the desired filter are designed in term of a solution to two coupled backward recursive Riccati equations. Finally, the effectiveness of the proposed filter design scheme is shown in an experimental platform.
This paper is concerned with the finite and infinite horizon optimal control issue for a class of networked control systems with stochastic communication protocols. Due to the limitation of networked bandwidth, only the limited number of sensors and actuators are allowed to get access to network mediums according to stochastic access protocols. A discrete-time Markov chain with a known transition probability matrix is employed to describe the scheduling behaviors of the stochastic access protocols, and the networked systems are modeled as a Markov jump system based on the augmenting technique. In such a framework, both the approaches of stochastic analysis and dynamic programming are utilized to derive the optimal control sequences satisfying the quadratic performance index. Moreover, the optimal controller gains are characterized by solving the solutions to coupled algebraic Riccati equations. Finally, a numerical example is provided to demonstrate the correctness and effectiveness of the proposed results.