Abstract
When a target moves in the field of interest (FOI), the availability of channel sight information is important since the appropriate use of such information can help to improve the localization accuracy. In this paper, Cramer-Rao Lower Bounds (CRLB) for a multi-sensor localization system under different channel sight conditions are computed analytically, which shows that significant degradation in accuracy can happen if channel sight conditions are misidentified. A novel received signal strength (RSS)-based localization algorithm is next proposed. In the algorithm, the FOI is assumed can be divided into distinct non-overlapped cells, and the cells experienced by the moving target are modeled as a hidden Markov model (HMM). Once the HMM matrices are obtained by off-line training, channel sight conditions can be identified by deploying a real time forward-only algorithm. These identified channel sight conditions are used to localize the target. To reduce the position estimation errors which result from the possible misidentified channel sight conditions, a relocalization algorithm is then used to review the identified channel sight conditions once abnormality in position estimate is detected. The simulation results show that our proposed localization strategy can provide good localization performance with the localization error approaching that when all channel sight conditions are known.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have