A Novel Framework for Energy-Aware Optimal Path Planning of Unmanned Aerial Vehicles

Source abstract

Unmanned Aerial Vehicles (UAVs) are becoming vital in many fields like robotics and aerospace. UAVs have many applications, ranging from delivering goods in disaster areas to agriculture. To make sure UAVs complete their mission efficiently, path planning becomes an important problem. The traditional models of path planning have constraints like finding the shortest path and obstacle avoidance. They ignore the fact that the UAV needs to have sufficient battery power to complete the mission or to stay connected to the ground station (GS). But in reality, this assumption of enough energy does not work. We address this gap by proposing an energy-aware model, SPSO-EA, by considering communication signal strength. The objective is to minimize a weighted cost function that includes path length, threat avoidance, altitude, smoothness, and communication energy, while satisfying obstacle avoidance and altitude boundary constraints. We divide the whole region into three zones depending on the signal strength. If a UAV navigates in a region with low signal, it ends up using more power to stay connected. By adding an energy-aware constraint, the proposed model works towards sustainable path planning. Experimental evaluation across 16 benchmark test cases using four optimization algorithms, PSO, GWO, DE, and CMA-ES, demonstrates that SPSO-EA consistently produces paths with superior energy efficiency. While the proposed model yields higher mean costs (e.g., 4924 vs. 4652 for PSO in Test Case 1), it significantly improves solution consistency, reducing the Coefficient of Variation from 7.4% to 4.8% in Test Case 3 and from 22.3% to 16.8% in Test Case 11. Among the four optimizers, DE with SPSO-EA achieves the lowest variance (CV as low as 4.7%), while PSO offers the best balance between solution quality and computational efficiency. These results confirm that the communication-aware energy constraint reliably guides UAVs toward energy-efficient paths regardless of the underlying optimization algorithm.