Home | About Me | Projects | Resume | Contact |
Date: Fall 2012
Disciplines: Robotics, Localization, Path Planning, Mechatronics, Inverse Kinematics Technologies Used: Python for high level planning code on laptop, Arduino for low level motor and arm control and serial communication to laptop, OptiTrack optical motion capture tracking system, TCP/UDP for position/orientation communication stream and diner/chef order updates to robot, Parallax EDDIE robot platform modified to use Arduino Team: Mark Jouppi, Omar Abdul-hadi, Hao Li, Yale Chen, Nishant Kumar. Advised by Dorsa Sadigh, Professors Lee and Seshia |
Description: Wouldn't it be cool if you could eat at a restaurant and be served by a completely autonomous robotic waiter system? To move towards this reality, my team and I designed and built a prototype robotic waiter system in a semester. Diners are seated at tables with fixed positions. Each diner table has a laptop with an interface through which diners can place orders. A chef station is also at a fixed position where orders are placed when finished and ready for delivery. The robot waiter's goal is to fetch and deliver orders throughout the dining room as efficiently as possible. We fix infrared markers to the robot and use Optitrack motion capture cameras around the room to get the precise location of the robot. As orders become ready, the robot optimally delivers them throughout the room, minimizing the total path distance travelled and breaking ties based on who ordered first for fairness. The path is dynamically updated as new orders come in. See the below demo videos and slides! In the demos, we forgo having actual dining tables and just label the floor with cardboard signs denoting the table number. We also use wood blocks instead of real food for simplicity. In the video demo, we use a laptop off-camera and enter diner orders which the video details in the captions. |
Overall System Demo |
Arm Demo |
Presenting the project to the class |
Our slides |
Back to Projects |