Movement Intelligence
for Bipedal Agents
Research from Taiji Motion on movement quality, measurement frameworks, and the deployment requirements for humanoid robots in human-facing environments. Papers are offered openly under CC BY 4.0 for use by the research community.
Deployment-Grade Motion: A Framework for Humanoid Robotics in Human-Facing Environments
Humanoid robots are moving from factories into care facilities, classrooms, retail floors, and homes — places where how a robot moves matters as much as whether it completes its task. The field has sophisticated benchmarks for task success, but no comparable standard for motion quality.
This paper proposes Deployment-Grade Motion as a measurable deployment requirement and introduces a three-layer framework — Performance, Communication, and Integration — grounded in the principle of awareness. It identifies the Four Hands Taiji curriculum as a distinctively well-suited reference system, and closes with an invitation to foundation model developers, humanoid OEMs, biomechanics researchers, and standards bodies to engage with and refine the framework.
Biomechanics-Driven Foundation Models: Bridging the Proprioceptive Gap
Current humanoid agents face a "Proprioceptive Gap" — the inability to move with proactive stability. This research proposes a framework for encoding Taiji's biomechanical principles as movement primitives, shifting robotic control from reactive sensing to master-level proactive balance.
Measurement Specification — Position Paper 02
A companion technical document that operationalizes the Deployment-Grade Motion framework into specific, reproducible measurement criteria — sensor requirements, threshold definitions, inter-rater reliability standards, and domain-specific benchmarks across Performance, Communication, and Integration layers. The first section (Smoothness, Performance Layer 4.1) is in active drafting.