Mobile robot path planning based on an improved Tornado Optimizer

Abstract Mobile robot path planning is a critical component of autonomous navigation. To address the issues of low search efficiency and susceptibility to local optima in classical path planning algorithms under complex scenarios, this paper proposes an improved Tornado Optimizer with Coriolis force (ITOC). First, in the population initialization phase, a good point set method is introduced, and prime numbers are generated according to the obstacle density in each column of the grid map, so as to enhance the diversity of the initial solutions. Second, in view of the functional differences among the three subgroups—storm, thunderstorm, and tornado—differentiated mutation trigger probabilities are designed to balance global exploration and local exploitation. Third, the following mutation operation is devised: using the line connecting the previously determined point in the preceding column and the target point, the intersection with the current column is computed; a Gaussian probability model centered at this intersection is then constructed for sampling. Meanwhile, according to the slope direction of the connecting line, an attenuation factor is introduced to reduce candidates that deviate from the target point, ensuring that the mutation consistently and efficiently approaches the target. Comparative experiments are conducted with six baseline algorithms, including the original Tornado Optimizer, an improved genetic algorithm, an improved ant colony algorithm, differential evolution (DE), particle swarm optimization (PSO), and the JAYA algorithm, under two different obstacle densities and on a benchmark map adopted from the literature. The results demonstrate that the proposed algorithm achieves superior performance in terms of path length and search efficiency. In particular, it obtains the same shortest path as Dijkstra''s algorithm in all 30 independent runs across all seven test maps.