A city full of lost passengers...

16th Annual Engineering Physics "Uber-Bots" Competition featured Autonomous Self-Driving Robots, capable of navigating through a mock city, searching for passengers and driving them to their destination

For das_pan, our design philosophy was to find the most basic, boring, and brilliant solutions to complex problems. The final result was a robot with large tolerances for sensor errors, robust electric circuits with reliable outputs, and a lightweight and flexible codebase for complex computations.

FEATURES

 Pickup Mechanism

Have you ever found yourself asking - if I were blind as a bot, how would I pick up a doll? We pondered the same question, and the obvious answer was by swatting at it. The width of the broom and pan allow for a large tolerance, meaning we don't have to be precisely lined up with the doll to pick it up.

 Seeing the World

Since computer vision is not feasible within 5 weeks, we used IR detection to traverse the map, follow tape, and find passengers. We liked IR detection so much, we even used it to encode the wheels on our robot, so that we could perform turns accurately. It took some signal processing, but das_pan was eventually able to see the world.

 It's what you're made of

Designed to withstand the force of stuffed toys, our chassis was cut from steel and welded together. Our pans and arm were actuated using RC servos, and our wheels were driven using motors. Fabrication involved the use of hand tools, laser cutter, water-jet cutter, 3D printer, and spot welder.

ROBOT

3D model

Try the demo below to move around and actuate a 3d model of das_pan.

Click to try