Path Planning for Autonomous Underwater Vehicles: An Ant Colony Algorithm Incorporating Alarm Pheromone
Yi-Ning Ma,Yue-Jiao Gong,Chu-Feng Xiao,Ying Gao,Jun Zhang
DOI: https://doi.org/10.1109/tvt.2018.2882130
IF: 6.8
2019-01-01
IEEE Transactions on Vehicular Technology
Abstract:Path planning is a critical issue to ensure the safety and reliability of the autonomous navigation system of the autonomous underwater vehicles (AUVs). Due to the nonlinearity and constraint issues, existing algorithms perform unsatisfactorily or even cannot find a feasible solution when facing large-scale problem spaces. This paper improves the path planning of AUVs in terms of both the path planning model and the optimization algorithm. The proposed model is comprehensive, which aggregates the length, energy consumption, and collision risk into the objective function and incorporates the steering window constraint. Based on the model, we develop a nature-inspired ant colony optimization algorithm to search the optimal path. Our algorithm is named alarm pheromone-assisted ant colony system (AP-ACS), since it incorporates the alarm pheromone in addition to the traditional guiding pheromone. The alarm pheromone alerts the ants to infeasible areas, which saves invalid search efforts and, thus, improves the search efficiency. Meanwhile, three heuristic measures are specifically designed to provide additional knowledge to the ants for path planning. In the experiments, different from the previous works that are tested on synthetic instances only, we implement an interface to retrieve the practical underwater environment data. AP-ACS and the compared algorithms are thus tested on several practical environments of different scales. The experimental results show that AP-ACS can effectively handle the constraints and outperforms the other algorithms in terms of accuracy, efficiency, and stability.
telecommunications,engineering, electrical & electronic,transportation science & technology