Abstract
In the aftermath of disasters such as bushfires, Unmanned Aerial Vehicles (UAVs) are essential for establishing communication and conducting search-and-rescue operations. However, navigating these complex, obstacle-rich environments requires efficient path-planning algorithms that can operate in real-time. This paper introduces a Cluster Optimization Scheme (COS) based on the Henry Gas Optimization (HGO) algorithm. By simulating the physics of gas solubility, HGO balances exploration and exploitation to find optimal flight paths. Experimental results demonstrate that HGO reduces transportation costs by 39.3% and computational time by 16.8% compared to industry-standard Particle Swarm Optimization (PSO).
Introduction
Disaster zones present unique challenges for autonomous navigation, including unpredictable terrain, rapidly changing conditions, and dense obstacles. Effective path planning must minimize distance while avoiding "local optima"—traps where an algorithm settles for a sub-optimal path because it cannot perceive a better global solution. Current industry standards, such as Particle Swarm Optimization (PSO), often suffer from instability or excessive computational demands in high-stakes scenarios where every second counts.
Methodology: Henry Gas Optimization (HGO)
The proposed approach utilizes the Henry Gas Optimization (HGO) metaheuristic, inspired by Henry’s Law regarding gas solubility in liquids. In this mathematical model, UAVs are represented as gas particles. Their "solubility" corresponds to the fitness of a specific trajectory (the shortest and safest path).
The algorithm operates through a structured process of initialization, clustering, and iterative updates. By dividing the UAV population into clusters with shared Henry coefficients, the system can simultaneously explore multiple regions of the search space. This clustering mechanism prevents the swarm from converging prematurely on a single, potentially sub-optimal solution, while the updating phase replaces the worst-performing paths to force re-exploration.
Experimental Results and Discussion
The HGO scheme was evaluated against four other metaheuristics—PSO, Grey Wolf Optimization (GWO), Cuckoo Search Algorithm (CSA), and Barnacles Mating Optimizer (BMO)—across four simulated environments in MATLAB: Ambient, Constrict, Tangled, and Complex.
Key performance metrics included:
- Path Efficiency: HGO achieved a 39.3% reduction in transportation costs in the Ambient scenario compared to PSO.
- Computational Speed: The algorithm was 16.8% faster than PSO, a critical factor for real-time disaster response.
- Global Optimization: In "Tangled" and "Complex" environments, HGO successfully avoided local optima that trapped other algorithms, consistently finding the shortest global path. For instance, in a complex scenario, HGO achieved a cost of 13.153 km, significantly outperforming CSA's 15.591 km.
Conclusion and Future Work
The HGO-based COS provides a robust and efficient solution for UAV trajectory optimization in disaster zones. By mimicking the stable search patterns of gas particles, the algorithm ensures faster and more reliable path planning. Future research will address hardware limitations, such as battery constraints and communication range, by exploring tethered UAV systems and LoRaWan integration to support real-time high-definition video transmission from disaster sites.
Summary updated