The Universal Rideshare Problem
The single most common cabin condition we see in real-world data is loose, dry debris: crumbs, sand, hair, receipts. It is the easiest task to demo and one of the hardest to fully solve at the boundaries — between the seat and the seatbelt anchor, in the seat-track gap, under the front edge of the floor mat. Our perception and tool design treats these edge cases as first-class, not afterthoughts.
Where Force Control Earns Its Keep
Sticky spills on cloth — soda, coffee with sugar, juice — are where vision-only approaches struggle. The visual signature is subtle once it dries; the mechanical signature (resistance against the wiper) is unmistakable. The force-controlled wiper described in the engineering posts is what turns this from a hard problem into a tractable one.
Streak-Free on Curved Surfaces
Modern windshields are aggressively curved. A wipe that works on a flat pane leaves streaks at the periphery of a windshield. The 6-DoF arm and impedance control let us maintain consistent tool orientation against the surface normal as the wiper traces the curve.
Deep Vacuum That Knows the Material
Carpet floor mats and rubber floor mats have very different cleaning profiles. The material classifier picks the right vacuum head and the right traversal pattern per mat type, and flags mats that should be removed and washed externally rather than cleaned in-place.
High-Touch Surfaces
Steering wheel, gear selector, door handles, infotainment touchscreen: a discrete set of high-touch points that benefit from a consistent disinfectant wipe between rides. We treat this as a sequenced sub-task with verification, not a free-form 'wipe everything' command.
Watch the latest demo reel
Visit handybot.ai →Hybrid position/force control for cleaning curved A-pillars without leaving streaks or scratching trim.