Fly-Hover-Communicate Protocol Overview

• Fly-Hover-Communicate (FHC) protocol is a structured approach that divides UAV operations into distinct fly, hover, and communicate phases for enhanced performance.
• It employs trajectory optimization, resource scheduling, and deep learning methods to improve energy savings and reduce mission time in various network scenarios.
• The protocol underpins applications from rotary-wing relays to intelligent transportation systems, achieving measurable gains in outage reduction and spectral efficiency.

The Fly-Hover-Communicate (FHC) protocol is a canonical operational paradigm for unmanned aerial vehicles (UAVs) used as communication relays, infrastructure nodes, or joint sensing–communication agents. It structures the mission into three temporally and spatially separated phases: (1) UAVs fly to designated locations; (2) hover at these points, performing communication or sensing tasks; (3) execute the data exchange or relaying operations, before resuming movement. This structure underpins a broad range of UAV networking designs—from rotary-wing communication relays and uplink NOMA systems to intelligent vehicular/traffic management and multiuser wireless networking—with mathematically grounded optimization frameworks for trajectory, time, and resource allocation, as surveyed in key works (Liang et al., 16 Dec 2025, Zeng et al., 2018, Zhou, 2014, Mozaffari et al., 2017, He et al., 2017, Mu et al., 2019).

1. Formal Definition and Structural Properties

The FHC protocol explicitly decomposes UAV operation into the following ordered steps:

• Fly: UAV departs from an initial position and follows a planned trajectory to a defined set of waypoints (task sites, communication zones, or clusters).
• Hover: At each waypoint, the UAV transitions to stationary or near-stationary flight, achieving fine positioning accuracy required for antenna alignment, imaging, or communications.
• Communicate (or sense): The UAV actively engages in data exchange (serving as a mobile base station, relay, or sensor platform) with ground users, infrastructure (e.g., RSUs, GBSs), or peer agents; resource allocation and scheduling are performed, communication links are dynamically established, and task-specific processing (e.g., relaying, traffic monitoring, or non-orthogonal access) takes place prior to flight resumption.

This structure is optimal in several system-theoretic senses: for example, the mission-time minimization in NOMA-based UAV uplink is characterized by a fly–hover–fly optimal trajectory, in which the UAV never communicates while flying slower than its maximum speed, and all communication is localized to designated hover-points (Mu et al., 2019).

2. Network and System Models

FHC designs arise in a variety of network topologies, all featuring mobility-constrained UAVs, energy, and communication resource optimization:

• Multi-hop or relay topologies: Several ground nodes (GNs) or users served by a UAV with restricted coverage, often with decode-and-forward relay modes (Zeng et al., 2018, Liang et al., 16 Dec 2025).
• Heterogeneous infrastructure: Integration with terrestrial RSUs, ground stations, or moving vehicles (e.g., in vehicular or railway scenarios) (Liang et al., 16 Dec 2025, Zhou, 2014).
• Single- and multi-UAV deployments: Task assignment, airspace partitioning, and spatio-temporal allocation for multiple cooperating UAVs, with non-trivial partitioning and resource coordination (Mozaffari et al., 2017, Liang et al., 16 Dec 2025).
• Sensing and communication fusion: Joint use of multi-view cameras, LiDAR, and onboard neural processing to support both physical sensing and network-side decision-making (Liang et al., 16 Dec 2025).

Channel models typically employ narrowband MIMO with AWGN, LoS-dominant air-to-ground propagation, and explicit consideration of interference management (NOMA/OMA, SIC), structured around achievable rate constraints, outage probabilities, and real-time handoff scheduling.

3. Mathematical Framework and Optimization

Central to FHC protocol design is the mathematical specification of trajectory, communication scheduling, and resource assignment:

• Trajectory optimization: Continuous or discretized paths linking stop points, governed by kinematic constraints ($\|\dot{\mathbf{q}}(t)\|\leq V_{\max}$), initial/final positions, and, if required, position-dependent coverage/feasibility sets (Zeng et al., 2018, Mu et al., 2019). The TSP (Travelling Salesman Problem) structure is prevalent for route ordering (Zeng et al., 2018, He et al., 2017).
• Hover time and energy trade-offs: At each hover point, the service time is determined by user throughput requirements and propagation geometry, for example:

$$T_k = \frac{\tilde Q_k}{B\log_2\left(1+\frac{\gamma_0}{H^2+\|\tilde{\mathbf{q}}_k-\mathbf{w}_k\|^2}\right)}.$$

The joint optimization is often non-convex; tractable reformulations employ convex approximation, path discretization, and successive convex approximation (SCA) methods (Zeng et al., 2018, Mu et al., 2019).

• Cell partitioning and load balancing: In multi-UAV deployments, optimal transport theory governs the spatial subdivision so that service cells $\mathcal{A}_i$ balance user density $f(x,y)$, bandwidth, and hover-time allocation (Mozaffari et al., 2017).
• Resource allocation: Bandwidth, power, and time-division scheduling, possibly with SDN-based control, is determined by per-user or per-cell requirements and fairness constraints (Mozaffari et al., 2017, Liang et al., 16 Dec 2025).
• Neural decision modules: In joint sensing/communication settings, feature extractors (e.g., multi-stream ResNet-18, 3D CNN), cross-agent attention (ACAF), and handoff/inspection heads form an integrated deep learning control pipeline (Liang et al., 16 Dec 2025).

4. Protocol Instantiations and Application Contexts

The FHC approach is the foundation of a spectrum of UAV applications:

• Rotary-wing UAV communication relays: Energy-efficient schemes for data collection or dissemination to ground nodes, with explicit analytic propulsion and communication energy models; demonstrable 20–40% energy savings over naive strategies, and additional gains with joint flying-while-communicating designs (Zeng et al., 2018).
• Airborne relaying for high-speed rail: Coordinated UAV handoff/association to high-speed trains with mmWave backhaul, leveraging trajectory smoothing, beam alignment, and fast control signaling to meet strict latency and coverage constraints (Zhou, 2014).
• Cellular NOMA uplink: Mission-time minimization via optimal fly–hover–fly trajectories subject to SIC, multi-cell interference, and per-user QoS constraints; solution via graph-theoretic sequences and SCA-based refinements yields up to 40% reduction in completion time against OMA (Mu et al., 2019).
• Integrated sensing/communication for ITS: Double-use UAVs for real-time relaying and concurrent multi-view traffic monitoring; UAP-Net combines learning-based fusion of vehicle and UAV sensors, proactive handoff control, and real-time inspection heads, outperforming both traditional beam sweep and LiDAR-only designs in outage and resilience metrics (Liang et al., 16 Dec 2025).
• Multiuser wireless networking with directional antennas: FHC protocol with altitude–beamwidth optimization, demonstrating that multicasting, broadcasting, and MAC modes yield substantially different optimal mission plans with regard to antenna parameters and flight geometry (He et al., 2017).
• Hover time minimization under fairness and load constraints: Optimal transport-based partitioning and per-user bandwidth allocation leads to closed-form minimum hover time and clear tradeoffs between spectral efficiency, cell size, and fairness (Mozaffari et al., 2017).

5. Key Mathematical Constructs

The following expressions are recurrent in FHC analysis:

• Point-to-point achievable rates:

$$R_{v,k} = \log_2\Bigl(1+\frac{|\mathbf h_{v,k}^H\mathbf w_{v,k}|^2}{\sum_{i\neq v}|\mathbf h_{v,k}^H\mathbf w_{i,k}|^2+\sigma^2}\Bigr),$$

$$R_{v,m,k} = \min\{\tau R_{m,k}, (1-\tau)R_{v,m}\}, \quad 0\leq\tau\leq 1.$$

• Fly–Hover–Communicate total energy:

$$E_{\rm tot} = E_0^*\sum_{k=0}^K\|\tilde{\mathbf q}_{\pi(k+1)}-\tilde{\mathbf q}_{\pi(k)}\|+\sum_{k=1}^K\frac{(P_h+P_c)\,\tilde Q_k}{B\log_2\!\bigl(1+\gamma_0/(H^2+\|\tilde{\mathbf q}_k-\mathbf w_k\|^2)\bigr)}$$

• Cross-agent fusion in neural FHC (UAP-Net):

$$\mathbf h^{\rm Coop} = \sum_{i} \alpha_i \mathbf V_i, \quad \alpha_i = \frac{\exp(\mathbf Q \mathbf K_i^T/\sqrt{L_c})}{\sum_j\exp(\mathbf Q \mathbf K_j^T/\sqrt{L_c})}$$

• Hover time with per-user loads (continuous, optimal transport):

$$\tau_i = \int_{\mathcal A_i} \frac{N\,u(x,y)}{B_i\,\log_2(1+\gamma_i(x,y))} f(x,y) dx\,dy + g_i\bigl(\int_{\mathcal A_i}f\bigr)$$

6. Performance Evaluation and Benchmarks

FHC-based designs are systematically benchmarked under various scenarios:

• Outage probability reduction: ~10 percentage point lower outage versus ground-only schemes at 200 Mbps requirement in vehicular relay/monitoring scenarios; robust resilience ($\rho\approx86\%$ with only one UAV present) (Liang et al., 16 Dec 2025).
• Energy efficiency: Significant (20–40%) reductions in total mission energy versus conventional approaches, further improved (by 15–25%) when hybrid flying-while-communicating is incorporated (Zeng et al., 2018).
• Mission completion time: Up to 40% improvement in NOMA setups relative to OMA, greater benefits at higher user data requirements (Mu et al., 2019).
• Monitoring and sensing: UAP-Net achieves comparable or superior detection error rates to state-of-the-art YOLO-based monitoring under adverse conditions (Liang et al., 16 Dec 2025).
• Bandwidth–hover time tradeoff: Increased bandwidth budget systematically reduces minimum hover times, but trade-offs exist with overall spectrum efficiency and coverage fairness (Mozaffari et al., 2017).

7. Summary Table: Representative FHC Protocols

Reference	Application Domain	Key Technical Innovation
(Liang et al., 16 Dec 2025)	Vehicular relay & sensing	UAP-Net: joint multi-modal fusion, proactive handoff
(Zeng et al., 2018)	Rotary-wing relay	TSP+convex site design, energy/mission time optimal
(Zhou, 2014)	Railway relay	Handover, mmWave beam, GPS/INS hover stabilization
(Mozaffari et al., 2017)	Multi-UAV area coverage	Optimal transport for fairness/load-driven partition
(He et al., 2017)	Multiuser comm. (alt/beams)	Altitude–beamwidth joint optimization
(Mu et al., 2019)	NOMA uplink, cellular	Fly–hover–fly proof, graph/SCA design, OMA–NOMA eval

These protocols collectively define the state-of-the-art in UAV communications subject to spatio-temporal mobility, power, and communication constraints, with the FHC structure as their unifying principle.