Autonomous Pick Up and Deliver Robot, University Project

Autonomous Pick Up and Deliver Robot, University Project

Tags
3D Printing
SolidWorks
Arduino
This robot was my favourite project university, and this course in general was a major factor in my decision to mechatronics. The course was built around a design challenge: build a small rover that would navigate a maze looking for a block, grab the block, and take it to a delivery zone.
My team! The pineapple was unrelated to the design challenge.
My team! The pineapple was unrelated to the design challenge.
I got to do a mix of CAD, mechanism design, electrical component selection and testing, and firmware development.
Diagram of the cad model showing the mechanical and structural components, actuator (servo), sensors (ultrasonic), and control devices (arduino and regulator).
Diagram of the cad model showing the mechanical and structural components, actuator (servo), sensors (ultrasonic), and control devices (arduino and regulator).
 
Another diagram with the more devices (IR, motor drivers, bigger servo, battery).
Another diagram with the more devices (IR, motor drivers, bigger servo, battery).
Finished build - this was before I learned about cable management…
Finished build - this was before I learned about cable management…
The robot used its ultrasonic sensors to determine where it was in a known map based on the wall topography. Those readings were transmitted to a Matlab script that estimated the location based on the sections of the maze it had seen. As the robot travelled further, it became more confident in its position and navigated its way to the known coordinates of the target block. While navigating, it used a PID feedback control system to keep a safe distance from the walls.
Flowchart describing the robot’s logic.
Flowchart describing the robot’s logic.
 
Here’s the final challenge video; there was a bonus for completing the challenge in 2 minutes, which we just barely missed. But we did get the bonus for no collisions! We probably focused a bit too much on collision avoidance and correctly predicting where we were in the maze, and should have spent a bit more time optimizing the speed. But it was a fun project and a fun course.