Background
Most edge computing research assumes a fixed topology of static servers. In reality, cities are in constant motion. Buses, trams, and trains carry passengers — and the devices they depend on — through coverage gaps, congestion hotspots, and unpredictable handoff zones. This project introduces Mobile Edge (ME) servers installed on public transport vehicles as first-class participants in the edge network, operating alongside Static Edge (SE) servers.
Each ME server is modelled as an intelligent, context-aware agent: it monitors its own compute load, its physical location and trajectory, the proximity of static nodes, and the service demands of passengers on board. As vehicles move through the city, these agents must continuously negotiate service handoffs — deciding when to accept a new service, when to offload to a static node, and when to collaborate with other vehicles in the vicinity.
Research Challenge
Future cities may increasingly rely on mobile computing infrastructure embedded within buses, trams, trains, autonomous vehicles and drones. These mobile edge platforms can bring computational resources closer to users, reducing latency and enabling new classes of AI-powered applications. However, unlike traditional edge servers, mobile platforms continuously change location, connectivity and resource availability, creating significant challenges for service deployment and management.
This project investigates how autonomous AI agents can manage service placement and orchestration across a hybrid network of Mobile Edge (ME) and Static Edge (SE) resources. Each mobile node acts as an intelligent agent, making local decisions based on its current workload, location, trajectory, connectivity and anticipated service demand. As vehicles move through the city, agents must continuously decide whether to host, migrate, replicate or terminate services while maintaining application performance and service continuity.
Key challenges include mobility-aware service placement, predictive service migration, agent coordination, intermittent connectivity, resource contention, and maintaining Quality of Service (QoS) as users and infrastructure move through the network. The project will explore whether techniques such as Reinforcement Learning and Multi-Agent Reinforcement Learning (MARL) can enable mobile edge agents to make effective autonomous decisions in highly dynamic environments.
Example application domains include AI-powered passenger assistants running on public transport, real-time augmented reality services for commuters, autonomous vehicle ecosystems, mobile environmental monitoring platforms, emergency response systems deployed on mobile units, and temporary AI infrastructure supporting large public events or disaster recovery operations. The work will involve developing and evaluating intelligent placement and orchestration strategies through simulation and experimentation.
Topics
Mobile Edge Computing (MEC) ; Autonomous AI Agents ; Multi-Agent Reinforcement Learning (MARL); Service Placement and Migration; Mobility-Aware Computing; Distributed Decision-Making; Edge AI and Autonomous Systems; Smart Transportation Infrastructure; Predictive Resource Management; Intelligent Service Orchestration
Impact
The future of urban computing may not be built solely on fixed infrastructure. Instead, cities could leverage fleets of connected vehicles, public transport systems and autonomous platforms as moving computational resources that deliver services wherever they are needed. This project explores how intelligent agents can transform mobile infrastructure into a dynamic computing fabric for future smart cities.
The outcomes of this research could support next-generation transportation systems, AI-powered public services, autonomous mobility platforms and resilient emergency-response infrastructures. More broadly, it addresses a fundamental challenge in distributed AI systems: how to deliver reliable intelligent services in environments where both users and computing resources are constantly on the move.