A Motion Planning System for Mobile Robots

Abstract

In this paper, a motion planning system for a mobile robot is proposed. Path planning tries to find a feasible path for mobile robots to move from a starting node to a target node in an environment with obstacles. A genetic algorithm is used to generate an optimal path by taking the advantage of its strong optimization ability. Mobile robot, obstacle and target localizations are realized by means of camera and image processing. A graphical user interface (GUI) is designed for the motion planning system that allows the user to interact with the robot system and to observe the robot environment. All the software components of the system are written in MATLAB that provides to use non-predefined accessories rather than the robot firmware has, to avoid confusing in C++ libraries of robot's proprietary software, to control the robot in detail and not to re-compile the programs frequently in real-time dynamic operations

A Modified Membrane-Inspired Algorithm Based on Particle Swarm Optimization for Mobile Robot Path Planning

“…We first discuss how to set the number m of elementary membranes in OLMS by using 20 × 20 grid model environment with 6, 8 and 10 obstacles, respectively. Then, 16 × 16 grid model environment with 9 static obstacles are applied to compare mMPSO with its counterpart vPSO and GA [15]. Subsequently, the complex environments, 32×32 and 64×64 grid model environments with 20 static obstacles, are applied to further test the mMPSO performance.…”

“…To investigate the mMPSO performance, this subsection uses three grid models, We first use the model with 16 × 16 grids to compare mMPSO with vPSO (when m = 1, mMPSO becomes vPSO) and GA in [15]. We consider three cases for K d , K s , K f as follows:…”

“…There are two ways of representing a grid-based environment. One is a X-Y coordinates plane [15] and the other is an orderly numbered grid, which has been widely used. We adopt the latter approach, in which a square environment is evenly divided into a certain number of squares, i.e., the x-axis and y-axis are divided equally into m parts, thus, we get m × m grids, where one or more grids are used to represent the obstacles.…”

“…Step 1 : An OLMS (6) Evaluate every particle (7) Find (8) Find local best particle (9) Execute communication rules (a) (10) Find global best particle (11) Execute communication rules (b) (12) Update particle's velocity V(t) (13) Update particle's position X(t) (14) Execute point repair algorithm (15) Execute smoothness algorithm (16) Adjust each particle's moving direction (17) Dissolve elementary membrane…”

“…The representative heuristic approaches for solving MR3P are neural networks, genetic algorithms [15], ant colony optimization, fuzzy logic [16], simulated annealing [17], PSO [21], probabilistic road maps, rapidly exploring random trees, etc. Although heuristic methods do not guarantee to find an optimal solution, they may be faster and may have higher efficiency than classical methods [19].…”