人工神经网络
水下
运动规划
路径(计算)
计算机科学
电流(流体)
工程类
人工智能
机器人
地质学
海洋学
电气工程
程序设计语言
出处
期刊:IEEE Transactions on Vehicular Technology
[Institute of Electrical and Electronics Engineers]
日期:2020-12-01
卷期号:69 (12): 14401-14412
被引量:44
标识
DOI:10.1109/tvt.2020.3034628
摘要
Path planning is a prerequisite for autonomous underwater vehicles to perform tasks autonomously. Many shortest distance algorithms are applied, and ocean currents are ignored to plan a short path in distance, which is usually time and energy consuming. In fact, the favourable currents can be exploited while avoiding the opposite ocean flows. Based on the bioinspired neural network architecture, this paper proposes a novel dynamic neural network model to plan the time-saving path in ocean current environments. After that, the path is smoothed by the B-spline algorithm. Analysis of the model shows that it can find out the minimum time path. Many simulations have also been introduced to test the effectiveness of the proposed model, showing good results. The dynamic neural network model has no learning procedure and can run in parallel. It has the advantages of loose parameter restrictions and wide spreading of neural activities. In addition, it has also been proven to be suitable for strong ocean currents.
科研通智能强力驱动
Strongly Powered by AbleSci AI