Please View Demo here: https://drive.google.com/file/d/1AtRoi2FTZ7aZQxGRw5f4dSL-5Zksno_m/view?usp=sharing

Team: Bartonium

Team: #17

Team Members: Prasham Yadothare @superiorcave, Eshaan Savanur @EstreemMC, Nachu Thenappan @NachuT

Description: Our group created a poker playing robot arm. This robot uses several cameras and a trained yoloV5 model to identify and determine the robot's hand per match. Then the robot decides to either check, raise, or match the oppenent based on its odds and the river (poker cards on the table). Using the SO-Arm101, the robot is able to draw its own cards and place "chips" during each of its matches. Using a waveshare Bus Servo Adapter and 6 FeeTech STS3215 servos the robot is able to move across the playing fields making several different actions. Using AMD's Cloud AI training GPUs, we were able to increase the accuracy of our yolov5 model exponetially and make our robot as accurate as possible.e.

Complete Project:

Additonal Pictures

• Competing aganist this model could help players train and rely on their hand get a grasp of poker without having to sacrifice any money. Additionally having a smart robot that is able to compute these decisions is just the start towards more intelligent automation that can complete smart tasks..

• Rather than using the standard libarary to control robot arm, our group used the open sourced servo adapter and servos in order to create our own programs to move the robot arm. We used the AMD AI cloud in order to train our yolov5 model for the opencv for the camera card detection. This is method of programming sets us apart and makes our project unique and innovative..

• Our model can be used in any environment as long as all of the objects (cards, chips, dispenser) are all place at the same relative distance from the robot arm.
• Our model is pretty adaptive as the model we trained is very accurate is and it could work in a variety of ways.
• To run this program, you need a computer connected via usb to control the arm and an esp32 that runs a websever and the dispenser we made.