Abstract:

The traditional ant colony optimization has some problems, such as weak initial path finding ability and slow algorithm convergence speed in solving the three-dimensional path planning problem of autonomous underwater vehicles. In this paper, a new method which combines Particle Swarm Optimization with improved Ant Colony Optimization（PSO-ACO） is proposed. The three-dimensional grid model is used to model the marine environment based on the idea of spatial stratification, and the path evaluation model is established by considering the length, ruggedness and danger of the path. Particle swarm optimization is firstly used for path pre-search to optimize the initial pheromone, and then the ant colony optimization is used to realize the global path planning, which improves the state transition rule, pheromone update method and adds the reward and punishment mechanism. The experiments show that the algorithm can effectively improve the ability of initial path finding and global searching, reduce the number of convergence iterations and shorten the search time.

Key words: improved ant colony optimization, particle swarm optimization, three-dimensional grid model, autonomous underwater vehicle, three-dimensional path planning

摘要：

针对传统蚁群算法在处理自主式水下机器人AUV（Autonomous Underwater Vehicle）三维路径规划问题时存在初期寻径能力弱、算法收敛速度慢等问题，提出一种融合粒子群与改进蚁群算法的AUV路径规划算法PSO-ACO（Particle Swarm Optimization-improved Ant Colony Optimization）。基于空间分层思想建立三维栅格模型实现水下环境建模；综合考虑路径长度、崎岖性、危险性等因素建立路径评价模型；先使用粒子群算法预搜索路径来优化蚁群算法的初始信息素；再对蚁群算法改进状态转移规则、信息素更新方式并加入奖惩机制实现全局路径规划。实验表明，算法能有效提高初期寻径能力和全局搜索能力，减少收敛迭代次数并缩短搜索使用时间。

关键词: 改进蚁群算法, 粒子群算法, 三维栅格模型, 自主式水下机器人, 三维路径规划