Vision Controlled Nios Robot Vi Co NBot Team
Vision Controlled Nios Robot Vi. Co. N-Bot™ Team Members Jeff Vickers (gte 613 i) Andre Moore (gt 6875 a) Kevin Walker (gte 143 x) K. Bosompem (gte 616 r) July 23, 2002 Georgia Institute of Technology School of Electrical and Computer Engineering
Overall Design Layout • The overall design layout helped structure the design process: • Phase 1: Research Phase 2: Nios & Camera Interface • Phase 3: Robot Controls Phase 4: Hardware & Power Design • Phase 5: State Machine & Testing
Overall Nios Schematic
Overall Software Design • Hierarchical structure of software implementation Including: ØGlobal Header File ØNios Design File ØLCD Module ØCamera UART Interface ØServo Speed & Direction ØState Machine
Pulse Width Modulation • A oscilloscope screenshot of a 2 ms pulse train with a 20 ms period generated using VHDL and integrated onto the Nios Board • A 2 ms pulse denotes full speed or right all the way
Pictures of the Vi. Co. N-Bot
State Machine • The state machine serves as a top level control to integrate all the Vi. Con-Bot functions • Integrates the CMUcam, Nios board, and hummer
Tweaking the Color Ranges • Testing was conducted in the Co. C and colors had to be recalibrated to accommodate the variation in color • Above is neon pink & green used to outline the track and denote when to negotiate a turn
New Color Measurements and Specifications • Above are the final color ranges • Both pink and green benefited greatly from recalibration
Track
Current Project Progress
Future Design Ideas • Implement addition sensors to allow for more advanced movements: sonar, multiple cameras, etc. • Improve state machine to include many different operations: follow an object, obstacle avoidance, search & destroy, etc. • Add a voltage sensor to detect the state of the battery • Design a swivel and tilt servo unit to pan the camera • Increase the clock frequency in the PWM to provide more precise resolution for important settings (home position) • Add a remote control for stopping robot manual at a distance • Utilize the next CMUcam soon to be released (faster & high res) • Tweak the camera interface to minimize propagation delays associated with serial communications in C & Nios • Add more color ranges for more advance tracking modes • Implement additional image processing • Use faster Nios processor
- Slides: 12