Source-linked AI summary
Joint Maneuver and Beamforming Design for UAV-Enabled Integrated Sensing and Communication
Zhonghao Lyu, Guangxu Zhu, Jie Xu
TL;DR
The paper addresses joint design of ISAC with a UAV deployment location or flight trajectory and transmit beamforming, while optimizing communication performance under sensing requirements. It develops SCA-, SDR-, and alternating-optimization-based algorithms, finding that proposed designs significantly outperform benchmark schemes and that sensing thresholds influence UAV placement and motion.
Problem
The paper studies how to jointly design ISAC with the UAV deployment location or flight trajectory and transmit beamforming while optimizing communication performance under sensing requirements.
Method
The paper develops high-quality and efficient solutions using SCA with SDR and 2D location search for quasi-stationary UAVs, and alternating optimization with SCA for mobile UAVs.
Results
The proposed designs significantly outperform other benchmark schemes, while higher sensing beampattern gain thresholds favor deployment near or flight toward the sensing area.
Takeaways & Limitations
The designs optimize data-rate throughput while ensuring sensing requirements across quasi-stationary and mobile UAV scenarios.
Abstract
from arXiv · showhide
This paper studies the UAV-enabled integrated sensing and communication (ISAC), in which UAVs are dispatched as aerial dual-functional access points (APs) for efficient ISAC. In particular, we consider a scenario with one UAV-AP equipped with a vertically placed uniform linear array (ULA), which sends combined information and sensing signals to communicate with multiple users and sense potential targets at interested areas on the ground simultaneously. Our objective is to jointly design the UAV maneuver with the transmit beamforming for optimizing the communication performance while ensuring the sensing requirements. First, we consider the quasi-stationary UAV scenario, in which the UAV is deployed at an optimizable location over the whole ISAC mission period. In this case, we jointly optimize the UAV deployment location, as well as the transmit information and sensing beamforming to maximize the weighted sum-rate throughput, subject to the sensing beampattern gain requirements and transmit power constraint. Although the above problem is non-convex, we find a high-quality solution by using the techniques of SCA and SDR, together with a 2D location search. Next, we consider the fully mobile UAV scenario, in which the UAV can fly over different locations during the ISAC mission period. In this case, we optimize the UAV flight trajectory, jointly with the transmit beamforming over time, to maximize the average weighted sum-rate throughput, subject to the sensing beampattern gain requirements and transmit power constraints as well as practical flight constraints. While the joint UAV trajectory and beamforming problem is more challenging to solve, we propose an efficient algorithm by adopting the alternating optimization together with SCA. Finally, numerical results are provided to validate the superiority of our proposed designs as compared to various benchmark schemes.
I. INTRODUCTION
The paper motivates UAV-enabled ISAC as a way to address terrestrial sensing limitations and studies joint UAV maneuver and beamforming design for communication and sensing. It develops solutions for quasi-stationary and fully mobile UAVs and shows that maneuvering helps balance sensing and communication performance.
- Research gap: Prior ISAC beamforming studies mainly address terrestrial MIMO networks, while UAV-enabled ISAC research remains at an early stage.The paper identifies joint design of ISAC, UAV location or trajectory, and multi-antenna beamforming as an appealing challenge.
- Motivation: Terrestrial ISAC sensing can be hindered by blocked line-of-sight links, clutter, and severe round-trip propagation loss for distant targets.These limitations arise from ground obstacles and limited terrestrial transmit power.
- Motivation: UAVs offer likely line-of-sight aerial-to-ground links and controllable deployment near sensing areas, making them suitable aerial ISAC platforms.They can operate as quasi-stationary or fully mobile access points, including in disaster or temporary hotspot scenarios.
- System and objective: The proposed system uses one UAV with a vertically placed ULA to transmit unified information and sensing signals to users and ground targets simultaneously.The design objective is to optimize communication performance while guaranteeing radar sensing requirements.
- Proposed designs: For a quasi-stationary UAV, the paper jointly optimizes deployment location and information and sensing beamforming for weighted sum-rate under power and sensing constraints.It obtains a high-quality solution using SCA, SDR, and a 2D location search.
- Proposed designs: For a fully mobile UAV, the paper jointly designs trajectory and time-varying beamforming under sensing, power, and practical flight constraints using alternating optimization with SCA.Numerical results validate the proposed designs against benchmark schemes.
- Results: High sensing beampattern thresholds favor UAV deployment or flight toward the sensing area, whereas low thresholds favor locations near but not exactly above users.The results show that deployment and trajectory design balance the inherent sensing–communication tradeoff.
II. SYSTEM MODEL
The system uses one UAV as a dual-functional aerial AP with a vertically placed ULA to communicate with multiple ground users and sense potential targets simultaneously. The model discretizes a finite mission period and jointly represents UAV positioning, communication, sensing signals, channels, received rates, and sensing locations.
- System Architecture: A single UAV equipped with a vertically placed ULA serves as an aerial dual-functional AP for downlink communication and radar sensing.The UAV communicates with K≥1 ground users while sensing potential ground targets at the same time.
- Antenna Configuration: The vertical ULA makes UAV orientation irrelevant to beam departure angles, simplifying trajectory design, whereas other antenna configurations could introduce orientation as an additional design variable.The orientation-related extension is left for future work.
- Mission and Mobility Model: The finite mission period is divided into N time slots, with sufficiently small slot duration so the UAV location is approximately unchanged within each slot.The UAV flies at a fixed altitude H while its horizontal location may vary over time.
- Transmit Signals: The UAV transmits user-specific information beams together with a dedicated sensing signal, while communication signals are also jointly exploited for sensing.The beamforming variables are the user beams {w_k[n]} and the sensing covariance-related signal s_0[n].
- Communication Model: Ground-user channels are modeled as line-of-sight links determined by UAV-user distance and steering vectors, with each user receiving desired signal, co-channel interference, sensing-signal interference, and noise.The LoS model is adopted for essential design insights and can be extended to channels with additional NLoS links such as Rician fading.
- Sensing Model: Sensing targets are represented by J predetermined ground locations, which may be uniformly sampled for detection or selected near known target positions for tracking.Larger J can improve sensing accuracy but increases computational complexity.
A. Problem Formulation
The paper formulates sensing-constrained weighted sum-rate maximization for both a quasi-stationary UAV and a fully mobile UAV. The designs jointly optimize location or trajectory with communication and sensing beamforming under sensing, power, and flight constraints.
- Quasi-Stationary UAV: In the quasi-stationary scenario, the UAV remains at one optimizable horizontal location throughout the mission and the objective is weighted sum-rate maximization.The information and sensing beamforming variables are optimized jointly with the fixed deployment location.
- Quasi-Stationary UAV: The quasi-stationary formulation requires sensing beampattern gain at each target location to exceed d^2(q,m_j)Γ while respecting the maximum transmit power.User weights α_k encode rate priorities, with larger weights assigning higher priority.
- Quasi-Stationary UAV: The quasi-stationary problem is difficult because its objective is non-concave and the sensing constraints are non-convex due to coupling between location and beamforming.The formulation jointly couples the UAV deployment location and transmit beamforming.
- Fully Mobile UAV: In the fully mobile scenario, the UAV trajectory and time-varying information and sensing beamforming are optimized for average weighted sum-rate throughput.The formulation includes sensing and power constraints at different time slots, plus initial, final, and speed-related flight constraints.
- Fully Mobile UAV: The mobile problem is more difficult than the quasi-stationary problem because it contains more UAV trajectory variables.Its objective uses the received SINR of each user at each time slot.
B. Feasibility Checking for (P1) and (P2)
Feasibility is checked by separating sensing and mobility conditions. The stationary case uses semidefinite optimization and location search, while the mobile case reduces trajectory feasibility to graph reachability under speed constraints.
- Quasi-Stationary Feasibility: Quasi-stationary feasibility is reduced to a sensing-only feasibility problem by setting all communication beamformers to zero.The resulting problem optimizes the sensing covariance and UAV location under the sensing and power constraints.
- Quasi-Stationary Feasibility: For a given UAV location, the sensing-only feasibility problem is a standard SDP solved with convex optimization tools, followed by a 2D search over the interested area.If at least one location makes the SDP feasible, the quasi-stationary problems are feasible.
- Mobile Feasibility: Mobile feasibility first identifies locations satisfying the sensing constraints, then checks whether a speed-constrained trajectory connects the prescribed initial and final locations.This converts the problem into a reachability test over feasible locations.
- Mobile Feasibility: The mobile reachability test constructs graph edges between feasible locations whose pairwise distance is below V_max and applies depth-first search from the initial location.The final location being in the reachable connected component is the feasibility condition.
- Mobile Feasibility: If the final location belongs to the connected component reached from the initial location, a feasible trajectory exists; otherwise the mobile formulation is infeasible.This conclusion applies to the feasibility problem and consequently to the mobile optimization problem.
IV. PROPOSED SOLUTION TO PROBLEM (P1) FOR QUASI-STATIONARY UAV SCENARIO
For the quasi-stationary UAV, the paper develops an SCA-and-SDR procedure for jointly optimizing beamforming and deployment location. The relaxed subproblems are convex, admit an optimal rank-one beamforming solution, and yield convergence through non-decreasing objective values.
- Solution Strategy: The proposed quasi-stationary solution uses SCA and SDR together with a 2D search over the UAV deployment location.The method addresses joint information and sensing beamforming optimization under a given location and then searches locations.
- Successive Convex Approximation: SCA replaces the non-concave objective with a concave approximation constructed iteratively from a first-order Taylor expansion.The resulting lower bound is used to form the SCA subproblem at iteration l.
- Semidefinite Relaxation: SDR relaxes the rank constraints, producing a convex SDP that can be solved optimally by tools such as CVX.The relaxation is applied after the SCA approximation.
- Rank-One Recovery: An optimal rank-one solution for the relaxed beamforming matrices always exists, so Gaussian randomization is unnecessary.The rank-one existence result is stated as Proposition 4.1.
- Convergence: Iteratively solving the relaxed SCA subproblems produces monotonically non-decreasing objective values and ensures convergence of the SCA-and-SDR algorithm.The convergence statement applies to the algorithm for solving the sensing-only beamforming subproblem and its associated formulation.
V. PROPOSED SOLUTION TO PROBLEM (P2) FOR MOBILE UAV SCENARIO
The mobile-UAV problem is solved by alternating between beamforming and trajectory optimization, using SCA to handle the resulting non-convex subproblems. A trust-region mechanism controls the trajectory approximation and supports convergence.
- Alternating optimization first updates information and sensing beamforming for a fixed trajectory, then updates the UAV trajectory using the new beamformers.
- With a fixed trajectory, beamforming optimization decouples across time slots and each subproblem has the form of the quasi-stationary design problem.
- The trajectory subproblem is highly non-convex because trajectory variables enter the steering vectors, so the method uses trust-region-based SCA.
- SCA linearizes the non-concave objective and convexifies non-convex constraints around the current trajectory, while trust-region constraints maintain approximation accuracy.
- The overall algorithm alternates beamforming and trajectory updates until the objective increase falls below a threshold; sufficiently small trust regions ensure convergence theoretically.
VI. NUMERICAL RESULTS
The numerical evaluation measures communication performance with sum rate in a 1 km × 1 km setting containing eight users and eighteen sensing sample locations. Unless otherwise stated, the simulations use the stated antenna, power, flight, channel, and weighting parameters.
- The simulation area is 1 km × 1 km and contains K = 8 ground users and J = 18 sensing sample locations.
- The UAV uses M = 12 antennas with half-wavelength spacing and a default sensing threshold Γ = 5e−5 (−43 dBm).
- The default flight altitude is H = 100 m, maximum horizontal speed is 30 m/s, and maximum transmit power is Pmax = 0.5 W.
- Communication performance is evaluated using sum rate, with noise power k = −110 dBm, reference channel gain β = −60 dB, and equal rate weights αk = 1.
A. Quasi-stationary UAV Scenario
In the quasi-stationary setting, the UAV location and information and sensing beamforming are optimized jointly, with benchmark designs used for comparison. Increasing the sensing threshold moves the UAV toward the sensing area and reduces communication sum rate.
- Benchmark schemes: The communication-only benchmark optimizes deployment and information beamforming for communication while ignoring sensing requirements.
- Benchmark schemes: The sensing-only benchmark maximizes the minimum distance-weighted beampattern gain by optimizing the UAV location and sensing covariance matrix.
- Deployment results: Two symmetric UAV deployment locations arise for each sensing threshold because sensing and communication locations are arranged symmetrically.
- Deployment results: As Γ increases from 0 to −37 dBm, the UAV moves closer to the sensing area and farther from users, decreasing communication sum rate.
- Joint ISAC design: The proposed ISAC design places the UAV between users and the sensing area and distributes beampattern gains more uniformly to balance both functions.
B. Mobile UAV Scenario
In the mobile setting, trajectory and beamforming designs are evaluated against sensing-only, communication-only, FHF, and SF baselines. The proposed ISAC design adapts UAV motion and radiated energy to balance sensing and communication, while improving with additional antennas.
- Mobile design and benchmarks: The mobile design alternates trajectory optimization with time-varying information and sensing beamforming under practical flight constraints.
- Convergence: The trust-region-based SCA convergence behavior reaches convergence in about 12 iterations.
- Trajectories: The sensing-only UAV flies toward the sensing-area center, whereas lower Γ values make the proposed ISAC trajectory move closer to users.
- Trajectories: At Γ = −70 dBm, the proposed trajectory resembles communication-only flight because information signals can satisfy the sensing requirements.
- Beampatterns: The proposed ISAC design radiates energy toward circles of different radii to cover sensing areas and users simultaneously.
- Multiuser communication: Users at distinct UAV distances can be served simultaneously because the vertically deployed ULA produces diverse angles of departure and less co-channel interference.
- Performance comparison: The proposed ISAC and FHF designs outperform SF; with more antennas, their average sum rates increase almost linearly while SF saturates.
VII. CONCLUSION
The paper develops joint UAV maneuver and transmit beamforming designs for UAV-enabled ISAC under quasi-stationary and fully mobile scenarios. Numerical results show that the proposed designs significantly outperform benchmark schemes, while extensions to other scenarios remain future work.
- The UAV acts as a dual-functional aerial AP that simultaneously communicates with multiple users and performs radar sensing toward an interested area.
- Under quasi-stationary and mobile UAV scenarios, the paper jointly designs UAV placement or trajectory with transmit information and sensing beamforming.
- Efficient algorithms based on convex and non-convex optimization techniques solve the highly non-convex sensing-constrained rate maximization problems.
- Numerical results show that the proposed designs significantly outperform other benchmark schemes.
- Extensions to other antenna configurations and multiple or swarm UAVs are identified as future work.
APPENDIX
The appendix constructs an alternative semidefinite-relaxation solution and verifies its feasibility, positive semidefiniteness, rank-one structure, and preservation of the objective value. It also derives the matrix representation used for the problem formulation.
- A new solution {W̄_k} and R̄_s is constructed from the obtained optimal solution and corresponding beamforming vectors.
- The constructed matrices satisfy positive-semidefiniteness and rank-one properties, while preserving the beampattern-gain and transmit-power constraints.
- Cauchy–Schwarz arguments verify that R̄_s is positive semidefinite, complementing the established properties of {W̄_k}.
- The objective value achieved by the constructed solution remains the same as that of the original solution, establishing optimality for (SDR5.l).
- For (P7), A(q[n],u_k) is defined as a(q[n],u_k)a^H(q[n],u_k), enabling the objective to be rewritten in matrix form.
- The entries and Hermitian structure of A(q[n],u_k) are derived, yielding equation (23) after substitution.