An Improved Ant Colony Algorithm based on Target-Driven Gravity Force

Addressing the issues of slow convergence, trapping into the locally optimal solutions, and blinding search in conventional ant colony algorithm (ACA) for mobile robot path-planning, we develop an improved ant colony algorithm based on target-driven gravity force (ACA-TDGF). The goal is quickly finding an optimal route in a grid map that connects the starting and ending points, avoiding encountering random obstacles. In the ACA-TDGF algorithm, the ants are classified as strong, medium, or weak according to the valid path length in each iteration, while the ant leaves the updated pheromone along the path. This novel pheromones updating strategy provides diversity in pheromone levels at the initial whilst improving the performance by selecting the strong class, reducing the probability of trapping into the locally optimal solutions. ACA-TDGF algorithm combines the gravitational function with the ant colony algorithm, replacing the first-order gravitational function as a higher-order exponential form so that the algorithm has a faster convergence speed and higher global search ability. Simulation results show superior performance in the tested environment.