Authors: Abhendra Pratap Singh, Riya Rani, Arpit Dwivedi, Nandini Sharma, Aakriti Sharma
Abstract: Cloud computing has historically been vital to the rapid advancement of multi-robot systems and autonomous cars for data processing, model training, and decision-making. However, the increasing demand for scalability, data privacy, and real-time responsiveness has exposed significant limitations of centralized cloud systems, such as excessive latency, bandwidth dependence, and security flaws. Federated Learning (FL) and Edge Computing (EC), which work together to provide decentralized and privacy-preserving intelligence, have become the focus of research in an effort to overcome these limitations. This paper addresses autonomous vehicles and multi-robots operated as edge nodes that train machine learning models locally on their own data, sharing just updates or model parameters with a central server instead of sending unprocessed information. This decentralized method greatly lowers communication costs, improves data secrecy, and facilitates real-time decision-making—all of which are crucial for operations that depend on safety. This paper’s contribution is to thoroughly analyze the issues including non-IID data dissemination, constrained computational and energy resources, and possible security risks, notwithstanding their benefits. In order to improve scalability, trust, and dependability, future research will combine block chain technology, 6G connectivity, and digital twin simulation. All things considered, the shift from cloud-centric computing to federated edge intelligence represents a critical advancement in the development of intelligent, safe, and effective autonomy in robotic ecosystems and next-generation automobiles
DOI: http://doi.org/10.5281/zenodo.17490951
