The fields of urban mobility, energy management, control systems, power system research, and building energy management are experiencing significant transformations with the adoption of artificial intelligence (AI), reinforcement learning, and decentralized technologies. A common theme among these areas is the focus on creating more sustainable, efficient, and resilient systems.
In urban mobility, researchers are utilizing feedback control and reinforcement learning to manage traffic flow, optimize parking management, and reduce pollution. Notable studies include the proposal of a novel feedback control system for incentivized suburban parking utilization and urban core traffic relief, as well as the development of hierarchical deep reinforcement learning approaches to optimize electric bus charging schedules.
The energy management sector is witnessing a significant shift towards AI-driven solutions, including the use of decentralized technologies to optimize energy management in microgrids and enhance the reliability of power electronics converters. A study on AI-based methodologies in energy management systems highlights the potential of self-healing microgrids, while another proposal integrates reliability requirements into the design framework of power electronics converters.
Control systems and reinforcement learning are also being applied to more complex and realistic applications, with a focus on energy efficiency, optimal performance, and automated environment design. The introduction of the LineFlow framework provides a standardized approach for simulating production lines and training reinforcement learning agents, while a general approach for automated reinforcement learning environment design using multi-objective optimization has shown promising results.
In the realm of power system research, the development of more efficient and sustainable energy systems is a key priority. Machine learning and AI techniques are being applied to power system security assessment, fault detection, and load forecasting, with notable papers introducing novel approaches to power system security assessment using multi-task learning and mitigating multi-stage cascading failures using deep reinforcement learning.
Lastly, the field of building energy management is moving towards more personalized and adaptive systems, leveraging Human-in-the-Loop AI and reinforcement learning to optimize HVAC systems and improve user experiences. A novel framework for optimizing HVAC performance using real-time user feedback and fluctuating electricity prices has been proposed, while a fully automated thermal detector-based traffic light system aims to enhance accessibility for users with disabilities.
Overall, these advances demonstrate a clear trajectory towards more sustainable, efficient, and resilient urban mobility and energy management systems, driven by the integration of AI, reinforcement learning, and decentralized technologies.
