Weekly Updates

Week 10 Update: Final Competition

Challenges:

• Optimizing gameplay performance.

Accomplishments:

• Updated website.
• Painted robot!
• Competed in final competition!

Plan for Next Week:

• Finish website and course requirements.

Media:

Week 9 Update: Seeding Competition

Challenges:

• Integration.
• Throwing past an actual adversarial opponent.

Accomplishments:

• Completed integration and have full autonomy working.
• Tuned gameplay code for seeding competition complete with both cooperative and adversarial modes.

Plan for Next Week:

• Paint robot.
• Further tuning of gameplay code to optimize competition performance.
• Work on website.

Media:

Week 8 Update: Check-Off Demo

Challenges:

• Single Ardunio Uno did not have enough I/O for our robot. Need to chain another Arduino into the system.
• Integration of gameplay code with physical system.
• Handling python and Arduino serial USB communication.
• Ball sensor unintentionally triggering due to ball loading servo.

Accomplishments:

• Constructed final integrated design. Includes belt-driven custom printed flywheels, capture net, mobile base, and elevated platform.
• Implemented code that successfully catches and returns the ball to the human player.

Plan for Next Week:

• More integration and tuning of gameplay code (including vision integration).
• Final wiring and clean up of final design (trim elevated platform to fit in zone, possibly paint, etc.)
• Establish radio link telecommunication.
• If time permits, incorporate HRI into the robot via an LED panel to display information for the human player.

Media:

Week 7 Update: Check-Off Demo

Challenges:

• Vibrations due to flywheel weight and imbalance.
• Shorted ESC because of conductive carbon fiber.

Accomplishments:

• Created CAD drawings and began construction of final integrated design.
• Demonstration of throwing mechanism with new flywheels and stronger motors.
• Integrated separate components in order to develop gameplay code.
• Verification of serial communication for motor control and game states.

Plan for Next Week:

• Testing of pinned direct-driven and belt-driven design for throwing mechanism.
• Finish construction of final integrated design (includes final wiring).
• Establish radio link telecommunication.
• Integration of game state and vision code.
• Begin final tuning for gameplay.

Media:

Week 6 Update: Check-Off Demo

Challenges:

• Running AR tag code on Raspberry Pi is slow.
• Integration.
• Full Communication between Raspberry Pi and Arduino.

Accomplishments:

• Successfully implemented AR tag detection feedback loop between vision computer and robot base controller.
• Faster I2C communication between Raspberry Pi and Arduino (replacing Serial for now)
• Upgraded base of robot to be metal and much more sturdy.
• Developed mechanism to vary flywheel distance.
• Designed and 3D printed custom flywheels to be used with larger brushless motors.
• Integrated all control code into one program for Pi and one program for Arduino.

Plan for Next Week:

• Update throwing mechanism with new 3D printed wheels and bigger brushless motors.
• Establish radio link telecommunication.
• Iron out lag between the vision controller and the robot base.
• Replacement of AprilTags with ArUco tags.
• Final wiring.

Media:

Week 5 Update: Mid-Term Demo

Challenges:

• Running AR tag code on Raspberry Pi is slow.
• Beach ball only launches to closest play zone. Need to move flywheels closer in next design iteration to address this issue.
• Full Communication between Raspberry Pi and Arduino.

Accomplishments:

• Successfully implemented AR tag detection code that runs on the Raspberry Pi.
• Serial communication between Raspberry Pi and Arduino.
• Built first integrated physical prototype of throwing mechanism and mobile base.
• Prepared working subsystems for mid term demonstration.

Plan for Next Week:

• Redesign throwing mechanism and mobile base to throw ball farther.
• Establish communication.
• Establish vision processor to robot controller communication to enable robot movement which reacts to the ball/AR tag detection.

Media:

Week 4 Update: Sensor Demo

Challenges:

• AR tag detection is sensitive to in-plane rotation. Running AR tag code on Raspberry Pi is slow.
• Translating orientation of omni-directional wheels and servos to translational motion was unclear at first.
• Issues tracking more than one color on the beach ball.

Accomplishments:

• Developed preliminary AR tag detection code.
• Received servos and omni-directional wheels and prototyped the mobile base.
• Integrated sensor with servo that will be used for object avoidance of mobile platform.
• Developed rough CAD drawings for throwing mechanism.
• Developed preliminary ball tracking code based on a single color (pink, blue, or orange).

Plan for Next Week:

• Continue AR tag detection development and integrate with the orientation of the mobile base.
• Decide on ball tracking method and continue development of ball tracking algorithms.
• Build physical prototype of throwing mechanism and integrate with mobile base in preparation for week 5 demonstration.
• Establish vision processor to robot controller communication to enable robot movement which reacts to the ball.

Media:

Week 3 Update: Actuator Demo

Challenges:

Initial Servo Throwing Arm Prototype

Our initial approach was to use a servo to engage our throwing arm. Unfortunately, we found that a single servo does not have enough power to throw the ball the required distance. Thus, we switched to using a flywheel design using two motors with wheels attached. We plan to re-purpose the servo to push the beach ball through the flywheels.

Flywheel Prototype

We used two quadcopter brushless motors and electronic speed controllers (ESCs) that we had on hand. We attached TekBot wheels to the motors and spun the wheels simultaneously in opposite directions to prototype the flywheel mechanism. The control of the motors was initially handled by a flight controller and we had difficulties using an Arduino. However, successful control by an Arduino was implemented after some debugging.

Accomplishments:

• We prototyped our initial design for a throwing arm and showed that the servo driven arm is not strong enough to launch the ball the required distance.
• We developed a second design using flywheels.
• We prototyped the flywheel design using motors and ESCs we had on hand.

Plan for Next Week:

• For the sensor demo next week, we will be using computer vision code to detect AR tags. This will run on the Raspberry Pi and provide basic sensing for person detection.
• Once the wheels arrive, we can begin prototyping the base and basic movement control.
• We will continue the development of our flywheel mechanism to run via Arduino commands for easier interfacing. We will also provide more detail on integration of the various throwing mechanism components.

Media: