Co-Creating Always-On Decentralized AI Infrastructure to Promote Physical Activities and Well-being in Green Spaces

This project aims to co-design and co-create a novel, always-on, low-cost, and energy-efficient computational infrastructure in open green spaces to promote physical activity, social connectedness, and well-being. By transforming green spaces into interactive environments, the project encourages movement and community engagement through the integration of AI, wearables, and gamified experiences. These environments will motivate individuals to stay active, connect with others, and enjoy their surroundings, improving both physical and mental health.

Leveraging serverless decentralized collaborative learning, the infrastructure will enable efficient resource sharing and AI model distribution, ensuring seamless communication between devices. Techniques like gossip protocols and federated learning will ensure the system is scalable, privacy-preserving, and energy-efficient, particularly in outdoor, resource-constrained environments.

The primary objectives are to develop a scalable, privacy-preserving AI infrastructure that operates efficiently in outdoor settings and to create decentralized AI and edge computing techniques that provide real-time personalized feedback to users, promoting physical activity and social engagement. By incorporating design and gamification, the project will deliver engaging user experiences, encouraging people to participate in and sustain healthy behaviours. The system will be tested and validated in collaboration with experts from sport science, art, and design to ensure user-centric and inclusive outcomes.

The broader goal is to democratize AI, making it accessible outside traditional labs and data centres. By utilizing federated learning, data processing will remain on the user’s device, ensuring privacy while reducing energy consumption. This approach supports sustainability goals like the UK's Net Zero 2050 by promoting low-energy computational methods.

The project is grounded in cutting-edge research in AI, focusing on federated learning, edge AI, and decentralized networks. It combines these advanced techniques to develop a low-energy, privacy-preserving infrastructure for real-time health interventions in public spaces.

Collaborating with experts in sports science, art, and design ensures the system is aligned with real-world health metrics, culturally relevant, and user-centric. This interdisciplinary approach will contribute to AI, health technology, and pervasive computing fields, offering a sustainable alternative to traditional cloud-based solutions. By integrating AI into green spaces, the project will create healthier, more connected communities while supporting global sustainability goals.