Researchers from the National Institute of Technology (NIT), Rourkela, have developed an AI-powered autonomous system for monitoring and cleaning solar panels, securing an Indian patent for the invention. The team includes assistant professor Arun Kumar, professor Bibhudatta Sahoo, and PhD graduates Lopamudra Hota and Biraja Prasad Nayak. The patent is titled ‘Federated Learning Based Autonomous System and Method for Monitoring and Cleaning Solar Plant’. The system addresses energy losses of up to 40% on large-scale solar farms caused by dust, bird droppings, and pollutants. The technology uses a Federated Learning (FL) framework for privacy-preserving monitoring, fault detection, and optimized cleaning recommendations, and is currently at Technology Readiness Level (TRL)-3. The system combines federated learning, edge computing, AI, autonomous cleaning, and predictive maintenance. Bibhudatta stated the system costs approximately 10% of the cost of existing systems. The team plans to develop a hardware prototype with IoT sensing and collaborate with government agencies and industry partners for field deployments, supporting the Government of India's National Solar Mission and Net Zero ambition.