Paula Louro

The increasing complexity of urban mobility demands scalable and adaptive traffic management solutions capable of operating in dynamic and decentralized environments. This work presents a distributed Optical Internet of Things (O-IoT) framework that integrates Visible Light Communication (VLC) with Deep Reinforcement Learning (DRL) for real-time traffic signal optimization. In the proposed architecture, each intersection operates as an autonomous DRL agent within a cooperative multi-agent system. Agents process locally sensed traffic data and coordinate their actions through VLC-based communication links, enabling low-latency and infrastructure-independent information exchange. This distributed approach enhances system robustness and eliminates the need for centralized control. The control policy is learned using the Proximal Policy Optimization (PPO) algorithm within the SUMO simulation environment, considering heterogeneous urban scenarios with varying traffic densities and topologies. Agents are trained using cooperative reward mechanisms that account for vehicle queue lengths, waiting times, and pedestrian flow. Simulation results over 200 training episodes demonstrate that the proposed framework significantly improves traffic efficiency compared to centralized and non-cooperative baselines, achieving up to 23% reduction in average queue length and 18% increase in pedestrian throughput. Furthermore, VLC-based inter-agent communication improves coordination and mitigates congestion propagation across intersections. These results highlight the potential of integrating optical wireless communication with intelligent multi-agent systems for next-generation traffic management. The proposed approach provides a scalable, adaptive, and resilient solution for smart city infrastructures operating under real-time constraints.

Modern airports are complex, dynamic infrastructures integrating terminals, gates, commercial areas, and operational assets, including passengers, Autonomous Guided Vehicles (AGVs), baggage carts, and ground support equipment. Ensuring safe, efficient, and resilient operations requires real-time coordination among these heterogeneous agents. This work presents a dual-layer AI-driven airport management system leveraging Visible Light Communication (VLC) and Multi-Agent Reinforcement Learning (MARL) to enhance operational efficiency, safety, and situational awareness in critical airport environments. The architecture separates airport operations into two functional layers: an outdoor layer for AGV traffic, aircraft taxiing, and gate assignments, and an indoor layer for passenger circulation, baggage handling, and commercial services. Vertical segmentation minimizes conflict zones between humans and autonomous systems, improving safety and reducing congestion. Autonomous Deep Reinforcement Learning (DRL) agents are deployed across both layers, making decentralized, real-time decisions based on local and neighboring traffic data. VLC-enabled luminaires transmit high-resolution spatial and navigational information using On-Off Keying (OOK) modulation and tetrachromatic LEDs, received by hybrid amorphous SiC optical sensors in a mesh topology. This enables precise localization and coordination of AGVs and pedestrians, independent of GPS. Adaptive queue/request/response mechanisms prioritize high-demand zones, resolve conflicts, and optimize flow management. Wayfinding algorithms dynamically route passengers to gates, check-in counters, baggage claim, and lounges, enhancing operational efficiency and passenger experience. Simulation and experimental results demonstrate reliable VLC communication, accurate localization, improved AGV throughput, reduced passenger wait times, and more efficient aircraft turnaround under cooperative MARL strategies. By embedding AI-driven agents within a VLC-enabled optical network, the framework provides a scalable, resilient, and intelligent solution for airport operations. This study highlights the potential of integrating multi-agent reinforcement learning with photonic technologies to support critical infrastructure management, improve safety, and enable adaptive, data-driven decision-making in increasingly automated airport environments.

Urban traffic networks demand sensing and control mechanisms capable of operating under fast, stochastic, and multimodal conditions. This work presents a communication-aware, topology-driven Multi-Agent Reinforcement Learning (MARL) framework enhanced by the Strategic Anti-Blocking Phase Adjustment (SAPA) module and supported by a Visible Light Communication (VLC) optical sensing layer. By leveraging existing LED streetlight and traffic-signal infrastructure, VLC provides low-latency, high-resolution photonic data acquisition, enabling intersections to function as distributed edge nodes that exchange real-time vehicle and pedestrian states through V2I, I2V, and P2I optical links. Each intersection operates as an autonomous DRL agent, coordinating with neighbors through VLC-enabled inter-agent communication. SAPA dynamically reallocates green times based on upstream queues and downstream storage capacity, preventing spillback along radial and circular corridors while preserving pedestrian fairness. The integrated VLC–MARL–SAPA system is evaluated using high-fidelity SUMO simulations, demonstrating significant reductions in queue lengths, halting events, and travel times, alongside increased throughput and multimodal stability. Results confirm that embedding physical photonic constraints directly into MARL decision-making yields more robust cooperation, faster convergence, and superior network-level performance compared to non-communicating or fixed-time baselines. Overall, this work introduces a scalable photonics-driven sensing and control architecture for next-generation intelligent transportation systems, illustrating how VLC-enabled state propagation and decentralized learning can enhance resilience, efficiency, and safety in complex urban environments.

Autonomous vehicles operating in complex indoor facilities require reliable position references, vehicle-state exchange, and infrastructure-assisted coordination, particularly where Global Navigation Satellite System coverage is unavailable. This paper presents an integrated Visible Light Communication architecture that uses ceiling-mounted RGBV luminaires, vehicle transceivers, and signaling units to support Lamp-to-Vehicle, Vehicle-to-Vehicle, Vehicle-to-Infrastructure, and Infrastructure-to-Vehicle links. Information is transmitted using On–Off Keying and structured 64-bit frames containing channel identification, position or identification data, timing, and link-dependent operational information. Representative experimental signals illustrate the transmission and recovery of vehicle and infrastructure messages under the considered laboratory conditions. An automated warehouse populated by automated guided vehicles (AGVs) is used as a representative dual-use application scenario. Two complementary simulation studies previously developed by the authors are summarized to demonstrate how VLC-provided information can support traffic coordination. Among signal cycles of 80, 100, and 120s, the 100s setting produced the lowest halting in the considered high-density scenario. A separate arrival-rate analysis identified the central crossing point as the principal bottleneck, with fluid operation at 250AGV/h, operation close to capacity at 500AGV/h, and overload at 750AGV/h. The proposed architecture provides a functional connection between optical information exchange and higher-level autonomous-vehicle coordination. It has potential application in GNSS-denied indoor logistics facilities where radio frequency communication may be constrained, including dual-use Security and Defense environments.

Congestion at shared intersections can limit the scalability of Automated Guided Vehicle (AGV) fleets in high-density warehouses. This study evaluates a Deep Q-Network (DQN) controller for adaptive intersection management in a SUMO-based warehouse model with five crossings, including the shared central crossing CP1. After one 200-episode training run, the DQN was evaluated in three seeded 6200 s simulations per demand scenario. Against a study-specific cyclic fixed-time controller at 250, 350, and 400 AGVs/h, the DQN achieved unweighted mean reductions of 62.9% in signal queue, 59.1% in average waiting time, and 18.0% in AGVs waiting to load, while loading/unloading occupancy changed by less than 1%. A matched 750 AGVs/h stress test showed that the benefit did not persist after saturation: at CP1, the mean signal queue increased from 13.15 to 30.68 AGVs and the mean number waiting to load from 9.64 to 15.93 AGVs. Sustained CP1 congestion began between 350 and 400 AGVs/h for the tested layout. Retained test outputs quantify test-level dispersion, but the single training run cannot estimate training variability. These results quantify the DQN traffic-control policy under ideal availability of SUMO state information and do not validate a VLC system. VLC is presented separately as a proposed future implementation layer; communication delay, packet loss, occlusion, and other physical-channel effects were not modelled.

Inefficient urban traffic management remains a critical challenge, as conventional signal controllers—built on fixed timing plans—cannot cope with the dynamic nature of modern city traffic. This study addresses this limitation by developing a decentralized MARL-based framework capable of coordinating five interconnected intersections as a unified traffic cell. Central to the proposed solution is the Strategic Anti-Blocking Phase Adjustment (SAPA) module, which enables intersections to autonomously modify phase durations in response to real-time traffic conditions. The framework is designed to handle heterogeneous demand patterns, with particular emphasis on arterial corridors connecting urban centers to peripheral zones. Integration of a Visible Light Communication (VLC) network allows continuous monitoring of key variables, including vehicle kinematics and pedestrian activity, feeding the agents with rich environmental feedback. Experimental evaluation confirms the effectiveness of the approach: the SAPA-augmented DQN achieves roughly 33% shorter vehicle queues and a ~70% reduction in pedestrian waiting counts relative to a standard DQN baseline. Remarkably, these gains bring the value-based method to a performance level comparable to MAPPO, a considerably more complex multi-agent policy optimization algorithm, establishing SAPA as an efficient and scalable enhancement for intelligent urban traffic control.

The rapid growth of urban vehicle and pedestrian flows has intensified congestion, delays, and safety concerns, underscoring the need for sustainable and intelligent traffic management in modern cities. Traditional centralized traffic signal control systems often face challenges of scalability, heterogeneity of traffic patterns, and limited real-time adaptability. To address these limitations, this study proposes a decentralized Multi-Agent Reinforcement Learning (MARL) framework for adaptive traffic signal control, where Deep Reinforcement Learning (DRL) agents are deployed at each intersection and trained on local conditions to enable real-time decision-making for both vehicles and pedestrians. A key innovation lies in the integration of Visible Light Communication (VLC), which leverages existing LED-based infrastructure in traffic lights, streetlights, and vehicles to provide high-capacity, low-latency, and energy-efficient data exchange, thereby enhancing each agent’s situational awareness while promoting infrastructure sustainability. The framework introduces a queue–request–response mechanism that dynamically adjusts signal phases, resolves conflicts between flows, and prioritizes urgent or emergency movements, ensuring equitable and safer mobility for all users. Validation through microscopic simulations in SUMO and preliminary real-world experiments demonstrates reductions in average waiting time, travel time, and queue lengths, along with improvements in pedestrian safety and energy efficiency. These results highlight the potential of MARL–VLC integration as a sustainable, resilient, and human-centered solution for next-generation urban traffic management.

Urban traffic congestion leads to daily delays, driven by outdated, rigid control systems. As vehicle numbers grow, fixed-phase signals struggle to adapt to real-time conditions. This work presents a decentralized Multi-Agent Reinforcement Learning (MARL) system to manage a traffic cell composed of five intersections, introducing the novel Strategic Anti-Blocking Phase Adjustment (SAPA) module, developed to enable dynamic phase time adjustments. The goal is to optimize arterial traffic flow by adapting strategies to different traffic generation patterns, simulating priority movements along circular or radial arterials, such as inbound or outbound city flows. The system aims to manage diverse scenarios within a cell, with the long-term goal of scaling to city-wide networks. A Visible Light Communication (VLC) infrastructure is integrated to support real-time data exchange between vehicles and infrastructure, capturing vehicle position, speed, and pedestrian presence at intersections. The system is evaluated through multiple performance metrics, showing promising results: reduced vehicle queues and waiting times, increased average speeds, and improved pedestrian safety and overall flow management. These outcomes demonstrate the system’s potential to deliver adaptive, intelligent traffic control for complex urban environments.

Airports are complex environments where efficient localization and intelligent traffic management are essential for ensuring smooth navigation and operational efficiency for both pedestrians and Autonomous Guided Vehicles (AGVs). This study presents an Artificial Intelligence (AI)-driven airport traffic management system that integrates Visible Light Communication (VLC), rerouting techniques, and adaptive reward mechanisms to optimize traffic flow, reduce congestion, and enhance safety. VLC-enabled luminaires serve as transmission points for location-specific guidance, forming a hybrid mesh network based on tetrachromatic LEDs with On-Off Keying (OOK) modulation and SiC optical receivers. AI agents, driven by Deep Reinforcement Learning (DRL), continuously analyze traffic conditions, apply adaptive rewards to improve decision-making, and dynamically reroute agents to balance traffic loads and avoid bottlenecks. Traffic states are encoded and processed through Q-learning algorithms, enabling intelligent phase activation and responsive control strategies. Simulation results confirm that the proposed system enables more balanced green time allocation, with reductions of up to 43% in vehicle-prioritized phases (e.g., Phase 1 at C1) to accommodate pedestrian flows. These adjustments lead to improved route planning, reduced halting times, and enhanced coordination between AGVs and pedestrian traffic across multiple intersections. Additionally, traffic flow responsiveness is preserved, with critical clearance phases maintaining stability or showing slight increases despite pedestrian prioritization. Simulation results confirm improved route planning, reduced halting times, and enhanced coordination between AGVs and pedestrian flows. The system also enables accurate indoor localization without relying on a Global Positioning System (GPS), supporting seamless movement and operational optimization. By combining VLC, adaptive AI models, and rerouting strategies, the proposed approach contributes to safer, more efficient, and human-centered airport mobility.

This study combines Visible Light Communication (VLC) and Artificial Intelligence (AI) to optimize traffic signal control, reduce congestion, and enhance safety. Utilizing existing road infrastructure, VLC technology transmits real-time data on vehicle and pedestrian positions, speeds, and queues. AI agents, powered by Deep Reinforcement Learning (DRL), process these data to manage traffic flows dynamically, applying anti-bottlenecking and rerouting techniques. A global agent coordinates local agents, enabling indirect communication and a unified DRL model that adjusts traffic light phases in real time using a queue/request/response system. A key focus of this work is the design of reward functions for standard and rerouting scenarios. In standard scenarios, the reward function prioritizes wide green bands for vehicles while penalizing pedestrian rule violations, balancing efficiency and safety. In rerouting scenarios, it dynamically prevents queuing spillovers at neighboring intersections, mitigating cascading congestion and ensuring safe, timely pedestrian crossings. Simulation experiments in the SUMO urban mobility simulator and real-world trials validate the system across diverse intersection types, including four-way crossings, T-intersections, and roundabouts. Results show significant reductions in vehicle and pedestrian waiting times, particularly in rerouting scenarios, demonstrating the system’s scalability and adaptability. By integrating VLC technology and AI-driven adaptive control, this approach achieves efficient, safe, and flexible traffic management. The proposed system addresses urban mobility challenges effectively, offering a robust solution to modern traffic demands while improving the travel experience for all road users.