A Heterogeneous Dual-Network Framework for Emergency Delivery UAVs: Communication Assurance and Path Planning Coordination

Abstract: Natural disasters often damage ground infrastructure, making unmanned aerial vehicles (UAVs) essential for emergency supply delivery. Yet safe operation in complex post-disaster environments requires reliable command-and-control (C2) links; link instability can cause loss of control, delay rescue, and trigger severe secondary harm. To provide continuous three-dimensional (3D) C2 coverage during dynamic missions, we propose a Heterogeneous Dual-Network Framework (HDNF) for safe and reliable emergency delivery. HDNF tightly couples an Emergency Communication Support Network (ECSN), formed by hovering UAV base stations, with a Delivery Path Network (DPN), formed by fast-moving delivery UAVs. The ECSN dynamically safeguards mission-critical flight corridors, while the DPN aligns trajectories with reliable coverage regions. We formulate a joint optimization problem over task assignment, 3D UAV-BS deployment, and DPN path planning to maximize end-to-end C2 reliability while minimizing UAV flight energy consumption and base-station deployment cost. To solve this computationally intractable NP-hard problem, we develop a layered strategy with three components: (i) a multi-layer C2 service model that overcomes 2D-metric limitations and aligns UAV-BS deployment with mission-critical 3D phases; (ii) a 3D coverage-aware multi-agent reinforcement learning algorithm that addresses the high-dimensional search space and improves both training efficiency and topology resilience; and (iii) a 3D communication-aware A* planner that jointly optimizes C2 quality and flight energy, mitigating trajectory--coverage mismatch and improving routing safety. Extensive simulations show that HDNF markedly improves C2 reliability, eliminates outages in critical phases, and sustains high task success rates while reducing hardware deployment cost.

• The paper introduces a dual-network framework optimizing UAV-BS deployment and UAV path planning to eliminate C2 outages across mission phases.
• It employs a multi-layer C2 service model and a custom MARL strategy with graph-theoretic reward shaping for robust and dynamic communication.
• Simulation results show 100% task success, a 20% reduction in required UAV-BSs, and up to 58% faster training convergence.

Heterogeneous Dual-Network Framework for UAV-Based Emergency Delivery

Motivation and Problem Setting

Natural disasters disrupt ground transport and terrestrial communication, necessitating UAV deployment for rapid emergency logistics. Reliable C2 links are critical for UAV safety throughout complex 3D mission profiles. Traditional UAV-BS deployment methods focus on static ground user coverage, neglecting the dynamic, multi-phase 3D C2 requirements of fast-moving delivery UAVs. Consequently, existing strategies result in C2 blind spots, particularly during takeoff/landing and high-altitude cruise phases. The core challenge addressed in this work is the simultaneous joint optimization of 3D UAV-BS placement and UAV path planning such that end-to-end C2 reliability is maximized, and energy and deployment costs are minimized, under realistic post-disaster constraints.

Proposed Framework: HDNF Architecture

The paper presents a Heterogeneous Dual-Network Framework (HDNF) for tightly coordinated UAV-based emergency delivery. HDNF encompasses:

• Emergency Communication Support Network (ECSN): Aerial base stations (UAV-BSs) create 3D C2 mesh coverage, dynamically safeguarding mission flight corridors.
• Delivery Path Network (DPN): Fast delivery UAVs with routes constrained to stay within robust C2 regions.

This architecture enables closed-loop coordination: ECSN deployments are informed by multi-phase delivery trajectories, and DPN paths are planned within strong C2 regions enforced by the deployed backbone.

Distinctly, HDNF is structured around a multi-layer C2 service model. Coverage and C2 service are stratified across terminal, vertical, and cruise corridor phases, each with specific vulnerabilities and communication demands.

Joint Optimization Formulation and Algorithmic Innovations

The joint optimization is cast as a mixed-integer nonlinear program over task assignments, UAV-BS 3D deployments, and delivery UAV 3D path planning. The optimization objective explicitly balances UAV-BS deployment cost, total UAV flight energy, and phase-coordinated C2 quality utility, constrained by communication and kinematic limits.

Directly solving the NP-hard coupled problem is intractable. The framework introduces several algorithmic advances:

• Multi-Layer C2 Service Model: Coverage demand is partitioned by phase, and a normalized capacity metric encapsulates region- and phase-specific C2 feasibility. This forms a reward-shaping signal for learning algorithms, eliminating phase-specific blind spots and aligning deployment with actual mission criticalities.
• 3D Coverage-Aware Multi-Agent RL Deployment: The UAV-BS topology is optimized via a custom MATD3 with PER, with a shared spatial encoder backbone for global environment awareness. Graph-theoretic algebraic connectivity is used for reward-shaping to maximize mesh robustness.
• 3D Communication-Aware A* Path Planning: Trajectory generation integrates both energy and SINR-based C2 feasibility, steering delivery UAVs along communication-resilient paths while minimizing detours.

Task assignment utilizes a sequential insertion VRPTW heuristic respecting energy and service window constraints.

Experimental Results and Numerical Performance

Comprehensive simulation results validate the proposed framework:

• C2 Reliability: HDNF eliminates all C2 outages across all mission phases and scales tested, unlike baselines which exhibit significant outage rates as disaster area size and SINR requirements increase.
• Resource Efficiency: HDNF sustains 100% task success rate while reducing the required number of UAV-BSs by approximately 20% compared to dense grid or conventional MARL-based deployments.
• Training Efficiency: The shared-backbone MARL architecture speeds up policy convergence by 39%–58% relative to baseline decentralized MARL, without loss in asymptotic performance.
• Scalability and Robustness: The dual-network design remains effective as disaster area expands to 5000 × 5000 m²—HDNF scales gracefully, while legacy schemes lose both robustness and efficiency.
• Delivery Efficiency: Communication-aware trajectory planning minimizes energy and time penalties otherwise imposed by C2 blind zones present in less coordinated strategies.

Strong claim: The paper explicitly demonstrates that HDNF achieves zero C2 outages and 100% task completion while requiring significantly fewer UAV-BSs than baselines, and that prior approaches routinely fail to satisfy critical vertical and cruise-phase C2 reliability at practical SINR requirements.

Theoretical and Practical Implications

The coordination of ECSN and DPN via multi-layer C2 modeling and RL-driven deployment provides a blueprint for mission-integrated wireless communications in dynamic, infrastructure-denied scenarios. This integrated approach supersedes the limitations encountered by decoupled network and trajectory optimization under complex, time-varying 3D mission constraints.

The practical impact is significant for disaster-response robotics, last-mile delivery, and emergency wireless systems. HDNF's hierarchical MARL decomposition and phase-stratified reward design are extensible to other A2A/A2G coordination scenarios, including heterogeneous UAV networks and collaborative swarms.

Future AI Directions

The work highlights several promising future avenues:

• Incorporation of real-time map updates and online learning to address non-stationarity in large-scale, adversarial or rapidly-evolving environments.
• Transfer of phase-oriented joint optimization to multi-agent mobile edge computing or sensor-actuator UAV networks under hybrid objectives (latency, load balancing, multi-stream C2).
• Close integration of model-based RL or graph neural policies for further efficiency and generalization across mission profiles and physical environments.
• Formal robustness analysis under adversarial communication or adversarial terrain profile perturbations.

Conclusion

The HDNF establishes an advanced paradigm for UAV-based emergency delivery through explicit dual-network coordination, robust RL-driven deployment, and communication-aware trajectory planning. By eliminating C2 blind spots and reducing resource consumption, HDNF offers a resilient, scalable, and mission-effective blueprint for deployment in post-disaster and infrastructure-denied conditions. The core algorithmic structure is well-positioned for future abstraction to broader AI coordination tasks in wireless robotic systems.

Reference: "A Heterogeneous Dual-Network Framework for Emergency Delivery UAVs: Communication Assurance and Path Planning Coordination" (2604.12501).