ABSTRACT: A wirelessly controlled robot that classifies the type of materials and gives a visual feedback to the user.

ROLE:The goal was to classify materials based on its nature. The materials were divided into Metals, Plastics and Rubber. The Pictorial dataset for metals, plastics and rubber were taken and the bounding boxes were drawn.

With the help of CUDA acceleration via google colab, The training was completed and a tensorflow model was obtained. The tensorflow model obtained is then converted into a tflite model as I wanted to run the scripts for controlling the motion of the robot as well as material classifiation on a single main board.

After overclocking the processor I was able to simultaneously run both scripts with acceptable performance. In order to improve performance, I would use a Coral Accelerator the next time around.

LANGUAGES USED:
Python, Java, Kotlin, Javascript, HTML, CSS