Project
# | Title | Team Members | TA | Documents | Sponsor |
---|---|---|---|---|---|
17 | Fast Low Cost Swarm Robots Honorable Mention |
Michael Bartmess Paul Ernst Peter Cork |
Dongwei Shi | design_document0.pdf final_paper0.pdf photo0.jpg presentation0.pptx proposal0.pdf video video video |
|
The project is to create a fleet of low-cost robots capable of moving quickly in a coordinated fashion. Each robot will be relatively small, smaller than a human fist, and the area of movement would be roughly the size of a table. Our goal would be to build 16 of them. Since the cost of each robot is multiplied by the size of the fleet, making a design that allows the robots to coordinate without using too many sensors is highly beneficial. Currently, swarm robots either are too slow or are too expensive. A prime example would be the Zooids project from Stanford, where the robots each use a high-quality IR sensor to decode their coordinates beamed from an expensive 3000Hz projector. This both limits the granularity of the positioning system and increases cost. Our method for coordinating the robots would be to use a 1080p webcam mounted overhead and use machine vision to identify the robots locations and orientation. The vision system would then send the current locations, orientations, and destination locations to the robot over WiFi. From there, the robots would utilize the information to move towards the destination location. The vision system will likely run off of a laptop and use a router to send the information over WiFi. Each robot would be equipped with an ESP8285 SoC which integrates a microcontroller with a WiFi chip, an antenna, motor controllers, two stepper motors for precise movement, and a battery with charging circuitry. The shell of the robot and PCB would both be designed by us. The robot would also feature some variety of a vision target on the top to assist the camera in identifying the robot and its location. |