Multi-Shape Compatiable Robotic Intake Mechanism

The 2023 season, Charged Up, asked alliances to collect cones and cubes from the floor and the substations and score them on a grid of nodes. The bottom row accepts either piece, but the middle and top rows are split into dedicated cone and cube nodes, so a robot that can only handle one piece gives up most of the scoring board.

The two pieces could hardly be less alike. The cone is rigid rubber, about 33 cm tall on an 8⅜ in square base, and weighs 1 lb 7 oz. The cube is purple PVC fabric inflated to about 9½ in face to face, weighs a couple of ounces, has rounded corners, and no flat surfaces to speak of. One is heavy and hard, the other is light and squishy, and they need completely different things from a gripper.

The mechanism ended up with two effective compartments sharing the same rollers. The cube is caught between the lower roller and the backing wall of the mechanism, so a soft object gets squeezed against a hard flat surface and cannot squirm out. The cone is caught between the two rollers instead, so a rigid object is pinched from both sides and driven inward rather than pushed away.

That split is the whole idea. Instead of finding one geometry that somehow suits both pieces, the intake gives each piece its own path through the same hardware, and the driver simply approaches the piece the right way. Two motors drive the rollers through pulley systems.

The hard part was material. Rubber and fabric behave differently against a spinning roller, and a compound that bites the cone well would either slip on the fabric or grab it so hard it deformed the cube out of shape. There was no way to reason this out on paper, so it was tested.

Prototype rigs were built out of wood, extrusion, and clamps, hand-fed with a real cone and a real cube, so a roller pair could be swapped and retested in minutes. Laser cutting produced the flat plates and side panels for those rigs, which meant a change in roller spacing or backing wall position was a new cut file rather than a rebuild, and 3D printing produced the roller hubs, spacers, and brackets that held each candidate in place. Roller compounds and durometers were cycled through, spacing between the rollers was adjusted, and the same two pieces were fed in over and over until a combination held both.

Each round of prototypes got closer to something that could live on a robot: the wood frames gave way to a proper frame, the roller spacing settled, and the backing wall took the shape it needed to hold the cube without catching the cone. The printed parts carried over from prototype to competition, the hubs, standoffs, and pulley mounts on the final intake are 3D printed, so the geometry that had been tuned on the bench went onto the robot directly instead of being redrawn for another process.

Testing moved onto the robot itself, picking pieces up off the carpet at driving speed, which is a different problem from feeding them in by hand. Approach angle, roller speed, and how the piece settled once captured all had to work while the robot was moving.

The finished intake sat on the front of 694's robot and picked up both game pieces off the floor for the whole season, without the driver having to think about which mechanism to use. Not carrying a second intake left the weight and volume budget for the rest of the robot.

The robot competed through the season with StuyPulse, the top FRC team in New York City, at world championship level.