Adaptive location correction and path re-planning based on error estimation method in underwater sensor networks

Due to the accuracy limitation of the underwater positioning system and the interference of the external envi-ronment, the actual trajectory of the autonomous underwater vehicle (AUV) will deviate from the predetermined path during the underwater mission, which cannot guarantee the mission completion and safety. To reduce navigation errors, this paper introduces forward-looking sonar (FLS) and proposes a path correction scheme based on error estimation of underwater sensor networks: Derive the position and attitude of the AUV by pro-cessing FLS images to obtain the relative orientation between the robot and the target; combine the node in-formation in the time-varying navigation map to calculate the error with the planned path; then fused with the a priori error of the underwater sensor network to build the error system model; completion of path re-planning based on the improved Bionic Neural Wave Network (BNWN) algorithm and time-varying navigation map; adjust the weight parameters during the iteration to make the drift error converge and finally complete the correction of the path. Simulation experiments prove that the path planned in this scheme can neutralize the drift errors from the algorithm level, and the AUV can complete the capture of dynamic and static targets with a small error when tracking this path.

Automated summary

This paper proposes a path correction scheme based on error estimation of underwater sensor networks to reduce navigation errors of autonomous underwater vehicles (AUVs). The scheme utilizes forward-looking sonar (FLS) and incorporates node information from a time-varying navigation map to calculate the planned path error, which is then corrected using the improved Bionic Neural Wave Network (BNWN) algorithm. Simulation experiments demonstrate that the planned path in this scheme can neutralize drift errors and enable AUVs to capture dynamic and static targets with minimal error.