Key points are not available for this paper at this time.
Abstract Mobile robots have become a focal point in various industries due to their advancements in intelligence and extensive application demands, with path planning technology serving as a fundamental support for achieving autonomous navigation. Although the traditional ant colony optimization (ACO) algorithm has been widely applied to path planning, it still suffers from the tendency to fall into local optima, insufficient global search capability, and the generation of excessive redundant nodes. To address this, a hybrid path planning algorithm is proposed in this paper. The algorithm generates an initial path using the adaptive step rapidly-exploring random tree (AS-RRT-Connect), which narrows the search space and provides a more accurate pheromone distribution for the improved ACO (IACO), thereby significantly improving the convergence speed of the algorithm. Additionally, four improvement strategies are designed, including an enhanced state transition rule, a path crossover strategy, a pheromone update method, and a local optimization strategy, to further enhance the path planning performance of differential drive mobile robots. Multi-environmental experiments conducted on the Matlab and Gazebo simulation platforms confirm that notable improvements in path smoothness, convergence speed, and planning stability are achieved by the modified algorithm. Compared to the traditional ACO algorithm, a 69.6% reduction in the number of turns has been achieved, the running time has been shortened by 52.94%, the standard deviation of path length has been decreased by 65.2%, and the average path length has been reduced by 20.1%. Studies on actual vehicles have demonstrated that the AS-RRT-Connect-IACO algorithm is capable of producing smooth, ideal routes and meeting the worldwide navigation needs of four-wheel differential drive carts in real-world scenarios.
Wu et al. (Thu,) studied this question.