• Computer Science > Networking and Internet Architecture [Submitted on 18 Feb 2026] Title:Multi-Agent Meta-Advisor for UAV Fleet Trajectory Design in Vehicular Networks View PDF HTML (experimental)Abstract:Future vehicular networks require continuous connectivity to serve highly mobile users in urban environments. • To mitigate the coverage limitations of fixed terrestrial macro base stations (MBS) under non line-of-sight (NLoS) conditions, fleets of unmanned aerial base stations (UABSs) can be deployed as aerial base stations, dynamically repositioning to track vehicular users and traffic hotspots in coordination with the terrestrial network. • This paper addresses cooperative multi-agent trajectory design under different service areas and takeoff configurations, where rapid and safe adaptation across scenarios is essential. • We formulate the problem as a multi-task decentralized partially observable Markov decision process and solve it using centralized training and decentralized execution with double dueling deep Q-network (3DQN), enabling online training for real-world deployments. • However, efficient exploration remains a bottleneck, with conventional strategies like $\epsilon$-greedy requiring careful tuning. • To overcome this, we propose the multi-agent meta-advisor with advisor override (MAMO).

Article Summaries:

  • Researchers have developed a multi‑agent meta‑advisor system (MAMO) to optimize the trajectories of unmanned aerial base station (UABS) fleets in urban vehicular networks. The approach models trajectory planning as a multi‑task decentralized partially observable Markov decision process and trains agents with a double‑dueling deep Q‑network. MAMO introduces a meta‑policy that guides exploration across different service areas and takeoff configurations, while allowing agents to override guidance when it misaligns with a specific scenario. Simulations in three realistic urban settings show faster convergence and higher returns than tuned ε‑greedy baselines, and demonstrate that UABS deployment markedly improves network coverage compared to terrestrial‑only solutions.

Sources: