Enhancing Localization of Mobile Robots in Distributed Sensor Environments for Reliable Proximity Service Applications

• Main Authors: Xingzhen Bai, Zhijing Zhang, Lu Liu, Xiaojun Zhai, John Panneerselvam, Leijiao Ge
• Format: Article
• Language: English
• Published: IEEE 2019-01-01
• Series: IEEE Access
• Subjects: Mobile robot localization; unscented Kalman filter; wireless sensor networks; packet loss; measurements compensation
• Online Access: https://ieeexplore.ieee.org/document/8642869/

Mobile robots can effectively coordinate information among sensor nodes in a distributed physical proximity. Accurately locating the mobile robots in such a distributed scenario is an essential requirement, such that the mobile robots can be instructed to coordinate with the appropriate sensor nodes. Packet loss is one of the prevailing issues on such wireless sensor network-based mobile robot localization applications. The packet loss might result from node failure, data transmission delay, and communication channel instability, which could significantly affect the transmission quality of the wireless signals. Such issues affect the localization accuracy of the mobile robot applications to an overwhelming margin, causing localization failures. To this end, this paper proposes an improved Unscented Kalman Filter-based localization algorithm to reduce the impacts of packet loss in the localization process. Rather than ignoring the missing measurements caused by packet loss, the proposed algorithm exploits the calculated measurement errors to estimate and compensate for the missing measurements. Some simulation experiments are conducted by subjecting the proposed algorithm with various packet loss rates, to evaluate its localization accuracy. The simulations demonstrate that the average localization error of the robot is 0.39 m when the packet loss rate is less than 90%, and the average running time of each iteration is 0.295 ms. The achieved results show that the proposed algorithm exhibits significant tolerance to packet loss while locating mobile robots in real-time, to achieve reliable localization accuracy and outperforms the existing UKF algorithm.