The research introduces a low-cost mobile manipulation system, Mobile ALOHA, that is bimanual and supports whole-body teleoperation. This system enhances the ALOHA system with a mobile base and a whole-body teleoperation interface. Using data collected with Mobile ALOHA, the researchers performed supervised behavior cloning and found that co-training with existing static ALOHA datasets boosts performance on mobile manipulation tasks. The system can complete complex tasks such as cooking, opening cabinets, calling and entering an elevator, and cleaning using a kitchen faucet. The study highlights the potential of imitation learning in developing generalist robots capable of performing complex tasks in everyday environments.
Publication date: 5 Jan 2024
Project Page: https://mobile-aloha.github.io
Paper: https://arxiv.org/pdf/2401.02117
