A Real-Time Path Planning Algorithm for AUV in Unknown Underwater Environment Based on Combining PSO and Waypoint Guidance.

It is a challengeable task to plan multi-objective optimization paths for autonomous underwater vehicles (AUVs) in an unknown environments, which involves reducing travel time, shortening path length, keeping navigation safety, and smoothing trajectory. To address the above challenges, a real-time path planning approach combining particle swarm optimization and waypoint guidance is proposed for AUV in unknown oceanic environments in this paper. In this algorithm, a multi-beam forward looking sonar (FLS) is utilized to detect obstacles and the output data of FLS are used to produce those obstacles' outlines (polygons). Particle swarm optimization is used to search for appropriate temporary waypoints, in which the optimization parameters of path planning are taken into account. Subsequently, an optimal path is automatically generated under the guidance of the destination and these temporary waypoints. Finally, three algorithms, including artificial potential field and genic algorithm, are adopted in the simulation experiments. The simulation results show that the proposed algorithm can generate the optimal paths compared with the other two algorithms.

An Adaptive Prediction Target Search Algorithm for Multi-AUVs in an Unknown 3D Environment.

For a target search of autonomous underwater vehicles (AUVs) in a completely unknown three-dimensional (3D) underwater environment, a multi-AUV collaborative target search algorithm based on adaptive prediction is proposed in this paper. The environmental information sensed by the forward-looking sonar is used to judge the current state of view, and the AUV system uses this environmental information to perform the target search task. If there is no target in the field of view, the AUV system will judge whether all sub-regions of the current layer have been searched or not. The next sub-region for searching is determined by the evaluation function and the task assignment strategy. If there are targets in the field of view, the evaluation function and the estimation function of the adaptive predictive optimization algorithm is used to estimate the location of the unknown target. At the same time, the algorithm also can reduce the positioning error caused by the noise of the sonar sensor. In this paper, the simulation results show that the proposed algorithm can not only deal with static targets and random dynamic interference target search tasks, but it can also perform target search tasks under some random AUV failure conditions. In this process, the underwater communication limits are also considered. Finally, simulation experiments indicate the high efficiency and great adaptability of the proposed algorithm.

A Real-Time Reaction Obstacle Avoidance Algorithm for Autonomous Underwater Vehicles in Unknown Environments.

A novel real-time reaction obstacle avoidance algorithm (RRA) is proposed for autonomous underwater vehicles (AUVs) that must adapt to unknown complex terrains, based on forward looking sonar (FLS). To accomplish this algorithm, obstacle avoidance rules are planned, and the RRA processes are split into five steps Introduction only lists 4 so AUVs can rapidly respond to various environment obstacles. The largest polar angle algorithm (LPAA) is designed to change detected obstacle's irregular outline into a convex polygon, which simplifies the obstacle avoidance process. A solution is designed to solve the trapping problem existing in U-shape obstacle avoidance by an outline memory algorithm. Finally, simulations in three unknown obstacle scenes are carried out to demonstrate the performance of this algorithm, where the obtained obstacle avoidance trajectories are safety, smooth and near-optimal.