Source-linked AI summary
Average Peak Age-of-Information Minimization in UAV-assisted IoT Networks
Mohamed A. Abd-Elmagid, Harpreet S. Dhillon
TL;DR
The paper asks how a UAV can preserve IoT information freshness when serving as a relay for a source-destination pair. It jointly optimizes flight trajectory, energy, and service times through a convergent iterative method for the non-convex PAoI problem. Results show significant gains from trajectory optimization when energy is limited or packets are large, while propulsion energy and online optimization remain outside the scope.
Problem
The paper addresses timely IoT status delivery when weak direct links, packet losses, and energy constraints threaten information freshness.
Method
The paper jointly optimizes UAV flight trajectory, source and UAV energy allocations, and packet service times using an iterative algorithm with analytical convergence guarantees.
Results
Trajectory optimization achieves significant PAoI gains especially when available energy is limited and/or update packets are large.
Takeaways & Limitations
Optimal UAV trajectory design can improve freshness, and the study quantifies energy requirements for achieving minimum average PAoI.
Abstract
from arXiv · showhide
Motivated by the need to ensure timely delivery of information (e.g., status updates) in the Internet-of-things (IoT) paradigm, this paper investigates the role of an Unmanned aerial vehicle (UAV) as a mobile relay to minimize the average Peak Age-of-information (PAoI) for a source-destination pair. For this setup, we formulate an optimization problem to jointly optimize the UAV's flight trajectory as well as energy and service time allocations for packet transmissions. In order to solve this non-convex problem, we propose an efficient iterative algorithm and establish its convergence analytically. Closed-form solutions for some sub-problems are also provided. One of the sub-problems we solve in this procedure is to jointly optimize the energy and service time allocations for a given trajectory of the UAV. This problem is of interest on its own right because in some cases we may not be able to alter the UAV's trajectory based on the locations of the IoT devices (especially when its primary mission is something else). Our numerical results quantify the gains that can be achieved by additionally optimizing the UAV's trajectory.
I. INTRODUCTION
The paper addresses information freshness in IoT networks by using UAVs as mobile relays when direct links are weak. It jointly designs UAV trajectories, energy allocation, and service times to minimize average PAoI.
- Motivation: IoT applications require fresh measurements because outdated data can cause erroneous decisions, while energy constraints increase packet losses.The problem is especially acute for far-off devices with poor direct links to destinations.
- Background: AoI is used to quantify information freshness at the destination node.Prior work studied AoI under different arrival, delivery-order, source, service-discipline, and reliability assumptions.
- Related Work: Prior UAV research considered deployment, coverage, rate, stochastic-geometry coverage, throughput-oriented trajectories, and hover time.These studies motivate applying UAV mobility to freshness optimization.
- Contributions: The paper formulates joint optimization of UAV flight trajectory, source and UAV energy allocations, and packet service times for average PAoI minimization.The optimization covers transmissions at both the source and UAV.
- Contributions: An efficient iterative algorithm is proposed for the non-convex problem, with analytical convergence guarantees and closed-form solutions for some sub-problems.The fixed-trajectory energy and service-time sub-problem is relevant when mission constraints limit trajectory changes.
II. SYSTEM MODEL
The system models a source sending status updates to a destination through a half-duplex UAV relay under just-in-time transmission. AoI evolution, packet service phases, channel gains, and UAV mobility define the model.
- System Setup: The UAV relay is used because the source-destination direct link is assumed weak, and AoI measures destination information freshness.The source may represent an IoT gateway forwarding measurements to a cellular base station through the UAV.
- AoI Evolution: Under just-in-time transmission, each new packet starts service after the preceding packet reaches the destination, producing AoI resets at packet completion times.The model uses N update packets and service durations d_i to characterize peak values.
- Transmission Model: Each packet service time is divided into an uplink phase from source to UAV and a downlink phase from UAV to destination.The UAV is single-antenna, half-duplex, fixed-height, and represented by phase-specific ground positions.
- Trajectory Model: The UAV trajectory is approximated by the sequence of ground positions used during each packet’s uplink and downlink phases.Initial and final positions are fixed by q1,1 = q0 and qN,2 = qf.
- Channel Model: Uplink and downlink channels are modeled as line-of-sight links with Doppler effects compensated for.Throughputs follow Shannon’s formula and depend on transmission energy, noise power, and the modulation-coding SNR gap.
A. Problem Formulation
The problem minimizes average PAoI by jointly selecting UAV positions, transmission energies, and service durations under throughput, energy, and mobility constraints. Variable coupling makes the formulation non-convex.
- Objective: The objective jointly optimizes the UAV trajectory and source-UAV energy allocations for update transmissions.The formulation targets the minimum average PAoI.
- Constraints: Source and UAV energy budgets constrain their total transmission energies, while packet-service constraints enforce successful uplink and downlink delivery.The packet size is represented through the required transmitted bits.
- Constraints: UAV mobility constraints restrict the phase-specific positions and their transitions.The formulation includes the UAV’s maximum movement limitations.
- Non-convexity: Although the objective is affine in service durations, coupling among durations, energies, and UAV positions makes the problem non-convex.This coupling occurs in the successful-transmission constraints.
B. Proposed Solution
The proposed solver alternates between optimizing energy/service allocations and updating the UAV trajectory, using convex formulations and successive convex optimization for the non-convex problem. The procedure provides closed-form or dual-based allocation solutions and is guaranteed to converge.
- The overall block-coordinate algorithm alternates allocation and trajectory updates until the fractional PAoI decrease falls below a threshold.Each trajectory update guarantees a non-increasing average PAoI, and the lower-bounded objective ensures convergence.
- Given a UAV trajectory, P2 jointly optimizes packet energy allocations and service times and is a convex optimization problem.Its convexity follows from the concavity of the rate perspective function and affine remaining constraints.
- When source energy is sufficient, each update packet can use its minimum feasible service time; otherwise, optimal allocations are obtained through a dual problem and sub-gradient updates.The sufficient-energy case uses Lemma 2, while the energy-limited case repeatedly updates the dual variable until convergence.
- The optimal service time decreases as the UAV approaches the source node, while the UAV’s maximum velocity imposes a lower bound on packet service time.The source-link channel quality affects achievable service time, but the mobility constraint restricts the minimum allowable duration.
- For fixed energy and service allocations, trajectory optimization maximizes the minimum throughput across all packet transmissions.The auxiliary variable represents the minimum achievable throughput, subject to uplink and downlink rate constraints.
- Successive convex optimization replaces rate constraints with first-order lower bounds, yielding convex QCQPs that can be solved efficiently by standard solvers.The lower-bounded rates are concave quadratic functions of the trajectory variables at each iteration.
IV. NUMERICAL RESULTS
The numerical results show that optimizing the UAV trajectory substantially matters for average PAoI when energy is limited or update packets are large, while PAoI eventually saturates beyond resource thresholds. The optimized trajectory adapts its movement between uplink and downlink phases to system parameters.
- Experimental setup: The numerical setup uses source and destination locations [−800, 800]T and [800, 800]T, respectively, with h = 100 meters and Vmax = 50 meters/sec.Unless otherwise specified, the experiments use N = 10, S = 1 Mbits, B = 1 MHz, and ES = EU = 1.25 joules.
- Trajectory-performance comparison: The optimized trajectory improves minimum achievable average PAoI relative to a straight trajectory, particularly with limited source/UAV energy or large update packets.The comparison uses the optimized trajectory from P1 against the straight trajectory Qs, whose performance is obtained by solving P2.
- Parameter impacts: Minimum average PAoI saturates when Es exceeds an energy threshold or S falls below a certain update-packet size.In these cases, the source and UAV have no remaining energy limitation, and the saturated value can be evaluated using Lemma 2 for trajectory Qs.
- Trajectory structure: The optimized trajectory’s inter-phase distances and directions depend mainly on source/UAV energy, update-packet size, and maximum UAV velocity.Figures 4 and 5 examine the resulting optimal trajectory under these parameter changes.
- Trajectory structure: As ES decreases or S increases, the UAV moves closer to the source during downlink and closer to the destination during uplink.This movement pattern is used to achieve minimum average PAoI under more demanding transmission conditions.
V. CONCLUSION
The paper formulates UAV-assisted PAoI minimization by jointly optimizing trajectory, energy, and service-time allocations. It solves the non-convex problem with a convergent iterative algorithm and reports trajectory gains and energy requirements, while leaving propulsion energy and online optimization as extensions.
- Conclusion: The optimization jointly minimizes overall PAoI through UAV trajectory, energy, and service-time allocation.
- Conclusion: The proposed iterative algorithm efficiently solves the non-convex problem and is guaranteed to converge.
- Conclusion: Optimizing the UAV trajectory yields significant gains when source/UAV energy is limited or update packets are large.The results also quantify the energy requirements at both the UAV and source needed to achieve minimum average PAoI.
- Extensions: The formulation includes only communication-related UAV energy, leaving propulsion energy incorporation and online trajectory optimization for future work.