Project
# | Title | Team Members | TA | Documents | Sponsor |
---|---|---|---|---|---|
19 | SCARA Drawing Robot |
Bingzhe Wei Chenghao Duan Tianhao Chi |
Dongwei Shi | design_document0.pdf final_paper0.pdf |
|
Members: Bingzhe Wei (bwei6) Tianhao Chi (tchi3) Chenghao Duan (cduan2) Title: SCARA Drawing Robot Description: We propose to develop a drawing robot based on an SCARA robot arm https://www.youtube.com/watch?v=vKD20BTkXhk. Overall system processing flow will be as follows: 1. User inputs image to image-processing program on PC. 2. Image processing with program on PC: a. Style transfer via deep neural networks b. Clustering of similar colors c. Pixel fill algorithm to convert to vector strokes 3. Vectors sent to microcontroller program via USB or similar. 4. Microcontroller program does inverse kinematics and commands motors as necessary. The combination of the SCARA design and stepper motors will enable a very stable and fast drawing platform, while the proposed image processing algorithms enable multiple, arbitrary styles and provide high-quality visual effects. We expect steppers will reduce the need for control. Proposed circuit will contain an USB-to-Serial converter IC, ATmega644 microcontroller, stepper motor controllers, and optical phototransistors for feedback control, as well as associated support circuitry. We use an ATmega644 with 4K of RAM, double that of ATmega328 as found in regular Arduinos to ensure enough capacity for inverse kinematics, while power will be supplied via a standard wall plug adapter that outputs 12V DC. Our team also has access to GPUs for training deep neural networks. Our team members have taken the following courses: ECE 470 - Robotics ECE 486 - Control System SE 423 - Mechatronics ECE 515 - Control Theory and Design ECE 547 - Topics in Image Processing - Deep Learning CS 598 PS - Machine Learning For Signal Processing |