The UW Robotics team is at it again. It’s not like we haven’t seen them consumed by robot-fever before. We’ve seen them build robots that float, fly, race, and fight each other, but this time… they’re taking it to Mars.
It was two years ago that the team decided to compete in the NASA Sample Return Robot Challenge, held each year at Worcester Polytechnic Institute in Massachusetts. The challenge involves building an autonomous robot to search and collect a variety of objects over an uneven, obstacle-ridden terrain without human intervention of any kind and then bring them back to home base within two hours.
As it turns out, this is a pretty complex problem to solve. The terrain is unknown, uneven, and filled with obstacles. The lighting varies, affecting the robot’s ability to “see,” and finding its own way back home in an 80 000 square-metre area is not an easy proposition. Add to that a handicap of not using any earth-bound sensors like a GPS or a compass and you have yourself an incredibly challenging task.
The University of Waterloo Robotics Team has participated in the contest two years running and, this year, their robot, affectionately named “Professor Marslander”, in honour of their faculty advisor Professor Steven Waslander, director of the Waterloo Autonomous Vehicle Laboratory, was ready to show the world what it was really capable of.
After a less-than-perfect performance at the competition last year, the team was more motivated than ever to deliver this time around. Fuelled by buckets of determination and coffee, the team spent months designing and building the four-wheeled robotic rover that was capable of collecting samples autonomously. Using a Velodyne laser scanner similar to the one used by Google’s self driving car, an array of high definition cameras, and enough lines of code to make even the most seasoned hacker’s stomach churn, the team put together their most capable and sophisticated robot to date.
The tension was high when the UW robot took the field for level one of the competition, where the robot must collect a single sample and return it to home base within 30 minutes. If that task was completed successfully, the bot would move on to level two, where up to ten samples would have be collected within a two hour time limit. After a nerve-wracking 20 minutes, where the robot found, collected, and returned the sampled to the home base, the team cheered triumphantly. Unfortunately, the sample from the robot was a few centimetres off from the correct position on the home platform, so it did not qualify to continue to Level 2. However, the robot won the People’s Choice award for most popular robot design, and the Technical Achievement award for sample detection and collection during Level 1.
The team is composed of 25 members, including undergraduate students in MME, ECE, SYDE, and graduate students from the Waterloo Autonomous Vehicles Lab (WAVELab), the Vision and Image Processing Lab (VIP Lab), and the Pattern Analysis and Machine Intelligence (PAMI) group. The team wishes to thank the Faculty of Engineering, the Department of Mechanical and Mechatronics Engineering, WEEF, the Dean of Engineering, and Clearpath Robotics for their amazing support.
Now that the competition is done, no one has any intentions of retiring. The team plans to continue working on the rover platform to make it their flagship competition robot as well as a platform for mobile robotics research. They believe they’re just getting started. As always, the UW Robotics team is looking for volunteers, so if you love tinkering with robots, hop on by to the team’s bay in the E5 Student Design Center to check them out.
Leave a Reply