Cubature Kalman Filtering for Nonlinear Systems With Energy Harvesting Sensors Under Probabilistic Quantization Effects

In this article, the problem of cubature Kalman filtering (CKF) is investigated for a class of nonlinear systems, which are equipped with energy harvesting sensors and subject to probabilistic quantizations. Due to the constraints of network bandwidth, measurement signals are quantized by a probabilistic quantization mechanism before they are transmitted through the communication network. Energy is harvested from the surrounding environment by sensors equipped with energy harvesters. The objective of this article is to design a novel cubature Kalman filter by taking into full account the effects of probabilistic quantizations and energy harvesting sensors based on the three-order spherical-radial cubature rule. By solving matrix difference equations, the upper bound of the filtering error covariance (FEC) is recursively computed and then minimized by constructing a proper filter gain. Moreover, the boundedness of the upper bound regarding the FEC is also discussed, and the monotonicity of the minimum upper bound in relation to the quantization level is further analyzed. The effectiveness of the proposed CKF algorithm is demonstrated through a simulation experiment focused on a target tracking scenario.