Drone networks optimize energy use and safety in risky zones
Threat-Aware Energy-Efficient Deployment for Dynamic UAV Networks: A Multi-Agent RL Approach
Information TheoryArtificial IntelligenceMachine Learning
Summary
Flying drones that act like flying cell towers need to stay safe and use energy well, especially in dangerous areas. The authors developed a method that first figures out where drones should be safely placed, then assigns drones to those spots to save energy, and finally uses an advanced AI approach to adjust their paths and power in real time. Their tests found no safety problems while using energy more efficiently and adapting quickly compared to other methods. This method also works well when conditions change or when there are more drones or different threats.
What this means in practice
- •For network operators: Deploy UAV base stations safely in hazardous environments while optimizing their energy use and connection quality.
- •For emergency response teams: Manage drone fleets in disaster zones to maintain communications with zero safety violations and adapt to changing conditions.
Authors
Faisal Al-Kamali, Hussein A. Ammar, Francois Chan, James H. Bayes, Yasser Gadallah, Mohamed H. Ahmed
Abstract
Ensuring operational safety in threat-prone environments remains a critical challenge for multi-UAV networks serving as aerial base stations. This paper proposes an efficient framework to maximize global energy efficiency (EE) while promoting safe operation through threat-aware clustering and reward-based safety enforcement. The proposed framework is executed in three steps. First, a threat-aware K-means (TAKM) algorithm determines the minimum required UAVs and computes safe initial placements. Second, an optimal matching stage assigns physical UAVs to these centroids to minimize energy expenditure. Third, a threat-aware multi-agent twin delayed deep deterministic policy gradient (MATD3) algorithm dynamically optimizes trajectories, power, and user associations. Simulation results show that the proposed framework achieves zero observed safety violations in the considered scenarios while achieving superior EE and faster convergence than other learning methods and non-clustering baselines. Compared to heuristic optimization, the proposed framework outperforms the greedy particle swarm optimization (GPSO) and achieves performance comparable to that of the optimized PSO (OPSO), while incurring significantly lower online deployment computational complexity. Furthermore, the proposed framework demonstrates effective generalization to unseen user distributions, large UAV fleets, and different threat geometries, while maintaining zero safety violations.