WASP-NEST Cloud Robotics
|WASP-NEST Cloud Robotics
|Obaidullah Zaland <firstname.lastname@example.org>
|2024-02-01 – 2024-08-01
Autonomous robotic systems (ARS) can grasp and manipulate objects in their environment in order to perform complex tasks. ARS have potential applications ranging from human care to improvements in the industrialized manufacturing chain. While approximately 3 million ARS have been deployed as of 2020, they do not benefit from collaborative learning and networking between robots yet, which results in the application of ARS in limited task-specific environments.
Recent advancements in machine learning, including computer vision and natural language processing, along with a boom of data-private phenomena of federated learning, help in enabling ARS to learn complex models for object grasping and manipulation collaboratively from each other's experience, with the added benefit of data privacy and model personalization.
In this interdisciplinary project spanning machine learning, robotics, cloud computing, and real-time control, we focus on achieving a breakthrough in the foundational algorithms, machine learning, and system design requirements for scaling networked, distributed, robotic manipulation systems to a large network of cloud-connected robots. We envision this network to collect very large-scale manipulation training data continuously and to learn from past experience using federated machine learning dynamically. Our approach will allow robots to balance between centralized machine learning in the cloud and local processing of information using the computational resources of each individual robot while incorporating real-time control and network bandwidth constraints.