Finally . . . a Chair That Falls Down and Picks Itself Up Again
The battery-operated chair, intended as a work of art, looks like an ordinary kitchen chair until it collapses into six pieces — four legs, a chair back, and a seat which houses the robotics. For a moment, the chair seems lifeless. Then, it begins to move, rolling first to one leg, then another, gathering itself, so to speak. If a leg lands in an inopportune position, the seat nudges it into the proper orientation. After attaching its limbs, the chair seat still rests on the ground.
Then, the chair slowly rises in a process designer Max Dean compares to a fawn gaining its footing (See it in action at roboticchair.com/documentation.php). Dean first thought of the chair in 1985. To build it, he teamed with artist and industrial designer Matt Donovan and Raffaello D’Andrea, a professor at ETH Zurich, Switzerland.
The biggest challenge, says Donovan, was developing a way to lift the 28-lb chair off the ground. He started with a 26-mm servomotor fitted with a gearhead that provided a 1,526:1 reduction ratio for a peak torque of 4.5 Nm. He added his own 9:1 gearbox for a total reduction ratio of 13,734:1.
“We solved the problem with sheer torque, flat out power,” says Donovan. The trade-off is a corresponding drop in speed. “Motors with enough torque to stand the chair up quickly would be too big and increase the weight so much we couldn’t get enough battery power to run them.”
The lifting problem solved, Donovan turned to the challenge of attaching the legs. He fitted each leg with a projecting rod pierced crosswise by a hole. The rods fit into corresponding brackets in each corner of the chair that link to the lifting mechanism.
The docking process involves two steps. First, the chassis (seat) maneuvers to slip the rod into the bracket. Then, a servomotor drives a pinion, the drive pin, through the holes to anchor the leg in place. The rods have a second hole drilled at 90° to the first so the legs can dock no matter which side they fall on. The chair collapses when the drive pins retract.
Retracting and installing the drive pins requires more speed than torque, so Donovan chose a 26-mm gearmotor with a 43:1 reduction ratio for up to 1.2 Nm of peak torque, placing one in each corner. The same size motor propels the chassis to find the legs and chair back. The chassis can even rise up on off-center wheels to drive off of any obstacle on which it has landed. That power comes courtesy of a servomotor with a 592:1 reduction ratio for up to 4.5 Nm of peak torque.
The chassis also attaches to the chair back and raises it to a vertical position. After maneuvering next to the chair back, a grabbing mechanism links to a bracket on the bottom of the chair back. For the sake of symmetry, Donovan chose a lower-torque gearmotor to run the drive pin that attaches the chair back. After the back is attached, a different mechanism raises it.
The robotic chair runs via a wireless link from a PC-based centralized controller. Because of the problem’s indeterminate nature, the chair operates open loop with input from an overhead vision system. The idea was to reduce hardware as much as possible, shifting complexity to the software.
Reprints and Licensing
© 2014 Penton Media Inc.