Improved Tornado Optimizer Matches Dijkstra Shortest Path in Robot Navigation Tests

An improved Tornado Optimizer incorporating Coriolis force mechanisms matched the shortest path identified by Dijkstra’s algorithm across thirty independent test runs on seven benchmark maps. Developed to enhance mobile robot path planning in autonomous navigation, the modified metaheuristic demonstrated superior search efficiency and reduced path lengths when evaluated against six established baseline algorithms under multiple obstacle configurations.

Autonomous navigation depends heavily on efficient path planning, yet conventional algorithms frequently encounter low search efficiency or become trapped in local optima when deployed in complex operating environments. To address these operational constraints, researchers Yuhang Cen and Xinsheng Lai designed the improved Tornado Optimizer with Coriolis force, designated as ITOC, which integrates targeted modifications into initial population sampling, subgroup mutation probabilities, and directional search operations.

During the population initialization phase, the algorithm incorporates a good point set method to establish well-distributed initial candidate trajectories across the search area. The procedure generates prime numbers calibrated specifically to the obstacle density measured within each column of the environmental grid map, thereby systematically increasing the diversity of initial solutions before iterative optimization commences.

To manage the transition between global exploration and local exploitation, ITOC establishes differentiated mutation trigger probabilities across three distinct functional sub-populations: storm, thunderstorm, and tornado. By tailoring the probability of mutation according to the specialized behavioral role assigned to each subgroup, the optimization framework prevents premature convergence while maintaining focused refinement near promising trajectory candidates.

The mutation mechanism computes an intersection point on the current grid column using a line connecting the target point with the preceding column’s selected coordinates. ITOC samples replacement candidates from a Gaussian probability distribution centered directly on this calculated intersection, while applying an attenuation factor tied to the slope direction of the connecting line to suppress candidates that deviate from the destination, guiding mutations consistently toward the target.

The researchers verified the algorithm through comparative benchmark experiments against six optimization baselines: the original Tornado Optimizer, an improved genetic algorithm, an improved ant colony algorithm, differential evolution, particle swarm optimization, and the JAYA algorithm. Evaluated across two obstacle densities and a literature-standard benchmark map, the algorithm demonstrated consistent convergence, securing the exact shortest path length in all independent test iterations.

Subject of Research: Mobile robot path planning using an improved Tornado Optimizer with Coriolis force

Article Title: Mobile robot path planning based on an improved Tornado Optimizer

Article References: Cen, Y., & Lai, X. (2026). Mobile robot path planning based on an improved Tornado Optimizer. Discover Applied Sciences. https://doi.org/10.1007/s42452-026-09789-2

Image Credits: AI Generated

DOI: 10.1007/s42452-026-09789-2

Keywords: mobile robot path planning, autonomous navigation, Tornado Optimizer, Coriolis force, ITOC algorithm, good point set method, prime numbers, obstacle density, grid map, Gaussian probability model, attenuation factor, Dijkstra algorithm, genetic algorithm, ant colony algorithm, differential evolution, particle swarm optimization, JAYA algorithm, global exploration, local exploitation