How Radar Complements Vision in Humanoid Perception

A sensing technology borrowed from cars and drones, for the situations where cameras alone fall short.

Radar measures distance and velocity by emitting radio waves and analyzing their reflection, working reliably through conditions that defeat cameras and even LiDAR to some degree — fog, dust, smoke, and low light among them.

Why it's used sparingly in humanoids today

Radar's spatial resolution is generally coarser than vision or LiDAR, so it's less useful for fine object recognition or manipulation. Its main value in robotics has been in situations closer to automotive use — detecting the presence and velocity of nearby obstacles reliably regardless of lighting or weather, which matters more for outdoor mobile robots than most humanoid use cases today.

Where it does show up

Some humanoid platforms intended for outdoor or industrial environments with poor visibility include a radar unit as a robustness layer specifically for basic obstacle presence and distance, kept deliberately simple and treated as a backup to richer vision-based perception rather than a primary sensor.