基于视觉定位的水下机器人无通信高精度编队技术研究

Research on High-precision Unmanned Underwater Vehicles Team Formation without Communication Based on Visual Positioning Technology

摘要:

水下机器人集群技术是目前水下机器人技术领域的发展热点之一。针对以往基于通信的水下机器人编队存在的编队精度低、队形保持困难等问题，提出了基于水下矢量光图案及视觉定位的水下集群编队方法，并通过水池试验分别验证了视觉定位以及水下密集编队的功能和相关指标。试验结果表明：水下视觉定位能够达到不大于 3%的定位精度和不小于 2 Hz 的定位频率，能够为水下机器人自主航行提供准确连续的控制输入。基于视觉定位的水下集群能够实现相互距离 10 m 以内的密集编队，编队精度不大于 10%，较以往基于通信的编队方法有较大提升。

关键词:水下机器人;集群;编队;视觉;定位

Abstract:

Unmanned underwater vehicles swarm technology is one of the current development hotspots in the field of Unmanned Underwater Vehicles. Aiming at the low formation precision and difficult formation keeping existing in communication-based unmanned underwater vehicles team formation，a method of underwater swarm formation based on underwater vector light pattern and visual positioning is proposed. The functions and related indexes of visual positioning and underwater intensive formation are verified by pool test. The test results show that underwater visual positioning can achieve a positioning accuracy of no more than 3% and positioning frequency of no less than 2 Hz，which can provide accurate and continuous control input for autonomous navigation of UUVs. swarm based on visual positioning can achieve intensive formations within the distance of 10 meters between each vehicle while the formation accuracy is no more than 10%，which is greatly improved compared with the previous communication-based formation methods.

Key words:unmanned underwater vehicle;swarm;team formation;vision;positioning