Frameowrk to perform Federated learning on robots.
Models from each robot are encrypted and sent to a central server, where they are combined to form an updated model. The model is then sent back to the robots where they will continue to be updated independently.
The Bulk Synchronous Parallel algorithm was used to update the models