Path Planning for Autonomous Navigation in Grid Environments Using Salp Swarm Algorithm and Genetic Algorithm

Abstract

Global path planning for autonomous navigation aims to generate safe, efficient, and collision-free routes, particularly in complex and dynamic environments where uncertainty and obstacles pose significant challenges. Among various environment modeling techniques, grid-based map representations are widely adopted due to their simplicity, scalability, and effectiveness in discretizing space for computational processing. This work presents a hybrid path planning approach that integrates the Salp Swarm Algorithm (SSA) with a Genetic Algorithm (GA) to enhance navigation performance in grid-based environments. The proposed method leverages the strong global exploration capability of SSA to initially identify promising and feasible paths across the search space. These candidate solutions are then utilized to initialize the Genetic Algorithm, which further refines the paths through evolutionary operations such as selection, crossover, and mutation. This cooperative integration enables a balanced trade-off between exploration and exploitation, improving convergence behavior and avoiding premature stagnation. As a result, the hybrid SSA-GA approach produces higher-quality paths in terms of reduced path length, improved smoothness, better convergence stability, and enhanced computational efficiency. The performance of the proposed method is evaluated on six planar grid maps with progressively increasing size and complexity to test robustness and scalability. Comparative analysis is conducted against well-known algorithms including A-star (A*), Ant Colony Optimization (ACO), Genetic Algorithm (GA), Salp Swarm Algorithm (SSA), Improved Sparrow Search Algorithm (ISpSA), Proposed Grey Wolf Optimization and Genetic Algorithm (GWO-GA) and proposed Salp Swarm Algoritjm and Genetic Algorithm (SSA-GA) . Experimental results demonstrate that the proposed SSA-GA method achieves a 100% success rate across all tested scenarios, consistently generating feasible solutions where several standalone algorithms fail, while maintaining competitive runtime and superior path optimality.