Performance Analysis of Path Planning Techniques Based on Potential Fields

Path planning is a research sub area of robotic which aims to solve the problem of conducting a robot from a specific position to a goal position in a given environment. The environment can be static or dynamic. This problem is considered complex when more than one robot is considered in a dynamic environment and it is necessary to control multiple robots in the same time. In this work, it is investigated some path planning techniques that can be applied to control multiple robots in soccer robot environment. This environment is very interesting due to fact as the robots as the goal (the ball) are dynamic objects in the environment. A comparative analysis of the algorithms, Oriented Potential Fields and Locally Oriented Potential Fields, that are used as path planners, are presented in relation to as number of iterations as convergence time. Further, taking in account that these algorithms require a numerical relaxation step per iteration, it is also shown how these algorithms can become faster, just by changing the initial conditions of the linear system solver.