Genetic algorithm based intelligent station-keeping for spherical underwater robot

Abstract

In this paper, a spherical underwater robot is proposed that uses six degree of freedom thrust allocator to determine the magnitude and the direction of thrust required for each water-jetted bilge pump thruster to create force and moment equilibrium. However, in order to ensure safe operation of the underwater robot it is equipped with redundant thruster configuration and therefore is over-actuated. Therefore, the choice of a particular solution for thrust allocation is found using an optimization process. In this work, the thrust allocation problem is formulated as an optimization problem, with an objective to minimize the total power consumption of the spherical underwater robot. A set of all physically realizable surge, sway and heave force for the ith thruster is called as its Attainable Thrust Region (ATR). Since the thrust force produced by each thruster is restricted to its ATR, the power consumption of underwater robot is determined by establishing a relationship between the power consumption and the thrust of an individual bilge pump. The formulated optimal thrust allocation problem is solved using Mincon (Sequential Quadratic Programming) and GA (Genetic Algorithm) optimization algorithm. The percentage savings in total power consumption for thruster system using the GA as compared to Mincon method is 33.82 %.