Free-Flying Cabin Robot: Fan Thrust Station-Keeping

Fly to waypoint

Control loop (PD)

Position gain Kp 1.4

Damping gain Kd 2.2

Max fan thrust 1.0 N

Speed

–

Dist. to target

–

Thrust use

–

Rack contacts

How it works

Free-flying cabin robots like NASA's Astrobee or the ISS's CIMON have no wheels — inside a pressurized module they hover on internal ducted-fan thrust, the same way a spacecraft in orbit has no road to push against. Six vectored fan nozzles push the body toward a target position while a PD (proportional–derivative) loop keeps it from overshooting and oscillating against the racks lining the corridor.

F = clamp( Kp·(target − pos) − Kd·vel , maxThrust )
pos += vel·dt ;  vel += (F/mass)·dt

Kp — how hard the fans pull toward the target. Too low and the robot drifts sluggishly; too high and it launches itself past the waypoint.
Kd — velocity damping. Too little and the overshoot rings back and forth for a long time; enough and it settles in one smooth glide.
Max fan thrust — the physical ceiling on any single correction. A weak fan set can't out-accelerate a strong disturbance, so large Kp gets clipped and the response slows down anyway.
Obstacle avoidance — when on, a repulsive field around each rack blends into the thrust command as the robot passes close, bending the path around it. Turn it off with a tight, straight-line gain and the direct route can clip a rack corner.

Real-world relevance: this is the same station-keeping problem every free-flying assistant robot solves continuously — hold a stable hover, glide to a new position, and never bump the hardware — using pure thrust, because in microgravity there's nothing to grip and brake against except the air a fan can push.

Cabin robot Target waypoint Equipment rack Fan thrust plume

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.