John Hugo Marcoux III - Systems Engineer
 

Autonomous Robot Line Painter

Home Design Hardware Results Thanks

Final Photos

Final picture of the finished robot

View of the top of the robot and the laser triangulation hardware

View of the back of the robot along with the paiting setup

Challenges Encountered

Due to our ambitious goals when starting the project, there were plenty of difficulties while implementing our design. These included accurate speed control of both motors simultaneously as well as implementing the laser triangulation. Although both motors are the same model, they have different characteristics. Each motor had different idle PWM signals, making them difficult to control together.
The biggest difficulty of all however was the laser triangulation. The basic concept is that we only had the angles between the beacons and the distance between the beacons. With only these four known values, it made solving for all of the other unknowns surprisingly difficult. After a few days, we were able to obtain a solution to the system of equations which could be implemented on our microprocessors, which had very little computing power. I would challenge anyone curious about the problem to try to solve it themselves, as our group still struggled after having probably 10 different engineers look at it. Seeing the solution was eye opening, however, and was one of the biggest hurdles we encountered throughout the entire project.

Possible Modifications

If we were to implement this design again, we would make quite a few changes, which would all revolve around having a larger budget. This would allow for a multitude of things, including more precise motor control, more precise laser triangulation readings, and a more effective overall design. Implementing a PID (proportional integral derivative) controller would have given us much better control over speed matching. On top of this, we would have added a LIDAR laser detection system, which would have given us not only more precise angle readings, but also range on these readings which would have made the triangulation calculations much easier. Lastly, we would have used a bucket of paint or chalk powered by an air compressor, allowing us to run the robot for much longer without having to change the can. This design would also have been gas powered, using an alternator for the electric power, giving us more power and making the overall design more efficient.