Adaptive Path Planning for Multi-Drone Systems Based on SFS, DPF, and Learning Based Refinement
Volume 11 , Issue 3 , August 2025 , Pages 10-29
Authors
Srwa 1 ; Prof. Dr. Amin Kakshar 2
1 Salahaddin University-College of Engineering-Software and Informatics Department
2 Salahaddin University-College of Engineering-Software and Informatics Department
10.17656/sjes.10194
Keywords
Abstract
Path planning is critical for multi-drone systems to let drones fly through different areas by minimizing travel time and avoiding all possible obstacles. The drone industry needs path-planning algorithms to develop efficient solutions, including monitoring functions, military, and surveillance. The existing algorithms, such as A*, Dijkstra’s, RRT, and ACO, fail to deliver satisfactory results while generating significant computational costs during real-time power consumption and dynamics avoidance attempts. The Adaptive Stochastic Fractal Algorithm integrates Stochastic Fractal Search for global pathfinding, Dynamic Potential Fields for real-time obstacle avoidance, and Learning Based Refinement using reinforcement learning for adaptive behaviour. The multi-drone systems use the ASF algorithm to minimize travel duration, prevent redundant drone movements, and enhance drone path smoothness without neglecting environmental adjustments. The combination of stochastic search with learning ability within SFS enables it to perform global and local path optimization at an appropriate level, thus making it suitable for real-time drone control. MATLAB provides a simulation of the effectiveness of the proposed ASF algorithm, which runs for environments with two and three dimensions. The results show that ASF creates routes requiring 12.5% less travel distance and 18% decreased energy consumption, resulting in a 20% shorter operational time than existing drone control methods. The updated algorithm provides optimal coordination capabilities for drones while managing their navigation system, making it suitable for complex, dynamic drone flight operations.