In rapidly evolving technologies, cloud computing and big data are integrated with robotics and automation systems to facilitate low latency offloading and real-time collaboration of distributed devices. Cloud robotics and automation system can handle unlimited computational storage resources overcome onboard resources constraints. Robotics and autonomous systems are data sensitive and require preservation, decentralized reliability, minimal communication, and on-device computation.

A high degree of autonomy and level of intelligence is achieved by deep learning models for robotics and automation system. However, deep learning models faces unavoidable difficulty in data privacy and security. Federated learning is capable of providing a privacy-preserving framework in robotics and automation. Federated learning provides decentralized and distributed training of learning models exceedingly without exploiting user data to the centralized server, communicating only local model updates.

Federated learning is effectively applied in decentralized multi-robot systems and distributed autonomous systems. In cloud robotics, federated learning promote efficient and secure robot learning for different types of autonomous system. Federated learning in robotics and automation is an emerging field, and its recent research focuses on multidisciplinary approaches. Future scopes that need to consider are optimization of communication, energy efficiency at the edge, personalized FL, and further privacy and security improvements.