It's an exciting time to be doing research in humanoid robotics. The following are some of the open problems CRΔL is focusing on:
Human-Robot-Environment Co-optimization (HERO). While necessary as a key component, physical AI alone is insufficient for autonomous humanoid robots, particularly when the environment is unstructured, messy, and complex. Going forwards, we will need to co-optimize the robot’s environment in parallel with its autonomy.
Integration of reasoning and control in a closed-loop feedback in open worlds. Without integration, robots can either reason without acting (like a chatbot) or act without reasoning (factory, warehouse robots). Without closed-loop feedback, we are restricted to “impressive”, but otherwise less-than-useful, open-loop rollouts (dancing, parkour etc.). And without the open world constraint, robots will never move out of the lab.
Solving SMGs with humanoid robots. SMGs with wheeled robots are much simpler since the action space is two-dimensional. Humanoid robots are not only high-dimensional, but also underactuated. Therefore, in addition to motion planning, one must also solve whole body control problems that arise in an SMG.