A method to evaluate autonomous performance for Unmanned Surface Vehicle

Abstract

Adjustable autonomy is an important mechanism in advanced robotic system. With this mechanism, robot can cooperate with operators, and the efficiency of mission is improved. In order to evaluate the advantages of the adjustable autonomy mechanism, a method of evaluating autonomous performance of Unmanned Surface Vehicle (USV) is proposed. This method considers two aspects in navigation task: the effectiveness of the task and the complexity of the environment. Several specific evaluation indicators are put forward, and the weights of indicators are determined by eigenvector method. Then a comprehensive evaluation value of task effectiveness and the environment complexity are calculated respectively. In two groups of the simulation experiment, USV performs navigation task in complicated marine environment. By comparing experimental data in different two modes, it can be seen that the performance of task in adjustable autonomy mode is better than that in full autonomy mode.