OSU Mars Rover
The goal of the project was to design and build a robot to compete in the annual University Rover Challenge which is put on by the Mars Society in Hanksville, Utah. The challenge seeks to develop a rover that would be representative of the type that would be used by humans after they've arrived on Mars (See here the official URC website: urc.marssociety.org). The competition is still fairly young, being first held in 2007, and is still evolving. Oregon State first competed in this competition in 2008. I joined the OSU Robotics Club In the spring of 2009, while still a student at Oregon State University, to participate in the Mars Rover project. I was part of the team for the following years, holding the following positions:
2009
Mechanical Engineer
2010
Lead Mechanical Engineer
2011
Project Manager
2008 (1st place)
A small group of Electrical and Computer Engineering students at OSU modified an existing robotic chassis (donated by a sponsor) to drive remotely using a wireless serial data link and analog video transmission. The team took 1st place in 2008.
2009 (4th place)
For the 2009 competition, the team committed to building a custom chassis of their own design. The concept was quite simple, and consisted of a sheet metal box frame that would house the motors, batteries, and video/data transmitters; antennas and cameras were mounted externally. I was asked to join the team to help resolve a camera stability issue. The pan-tilt system then in place was not stiff enough, and allowed the camera to spring back and forth as the vehicle traversed rough terrain. Because the camera was mounted relatively high, accelerations were of similar magnitude in all directions. I elected to build a fully spring support for the pan-tilt system using RC shock absorbers (similar to the concept of a Stewart Platform). The system provided significant travel but unfortunately the total inertia of the sprung mass was insufficient to provide a stable platform for the shocks available, and provided only a marginal improvement in video quality.The team took 4th place in the 2009 competition. The design had suffered from a number of obvious flaws that left several competition tasks incomplete, some of which are listed here:
Low ground clearance
In competition one of the tasks was left incomplete when the vehicle high-centered itself on a mound of dirt.
Highly sensitive speed control
In competition, the vehicle drove into a mock control panel while using its robotic arm, destroying the arms servos and prematurely ending the task.
Narrow tires with a tank style drive and chain and sprocket drive train
The heavy weight of the vehicle (110 lbs), coupled with the narrow tires and soft sandy ground, caused the vehicle to dig into the ground when turning in place. This resulted in a rapid increase in required torque which everything but the chain drive could provide, eventually causing a catastrophic failure of the drive train, leaving the vehicle stranded.
2010 (1st place)
There was much learned from the 2009 competition; unfortunately primarily in the category of how not to design such a vehicle. New designs were already being drawn-up in the van during the 17-hour return trip from Hanksville, Utah to Corvallis, Oregon. Effort would be made to avoid drive trains and power each wheel directly with its own motor, avoid narrow tires, and avoid long wire harnesses by strategically locating electronics in the chassis. I developed a chassis concept that could satisfy these requirements, and after constructing a scale Lego chassis demonstrator, the team agreed on the concept and I was elected lead mechanical engineer, in charge of all mechanical design, production, and validation. Several members of the team elected to take ownership of the design of the robotic arm as their senior year capstone project. The team agreed to this and the robotic arm became a separate electromechanical design project in parallel with the rest of the vehicle development.
Here is a picture of all the chassis components of the 2010 vehicle. All aluminum components in view were custom designed and machined by members on the team. Notice the motor controllers housed in the chassis structure; this both allowed the stock heatsinks to be removed, and controllers to be located far from more sensitive circuits as their high-current switching is a source of electromagnetic interference.
The 2010 design was incredibly successful (shown here in the Utah desert during the 2010 University Rover Challenge sample return task), placing 1st far and ahead of the 2nd and 3rd place teams! Part of the competition's 1st prize was to be invited to present at the 2010 Mars Society Conference. At the conference, which was held in Dayton, Ohio, I presented a thorough overview of the design process, decisions made, as well as an overview of the 2010 University Rover Challenge.
In addition to the items explained here, please take a look at the 2010 OSU Mars Rover Design Report as well as a 3D PDF file of the CAD model linked here below (must have adobe acrobat installed to view 3D PDF) which provides a good summary of the vehicle, it's systems, as well as team structure.
2010 Oregon State University Mars Rover Design Report.pdf
2011 (3rd place)
Based on the apparent recipe for success, the team approached the design of the 2011 rover much the same as the year before. Since so much of the new architecture seemed to perform so well, there was a much greater focus on making improvements to the design rather than drastic changes in architecture. Following my experience as Lead Mechanical Engineer, I volunteered and was elected as Project Manager. An emphasis was put on optimizing the chassis design to save weight. the 2010 rover's lead robotic arm engineer Josef Hortnagl was promoted to Lead Mechanical Engineer for the vehicle, and would spearhead the effort to cut weight from the chassis, the results of which are shown in the picture below. Once again the robotic arm design was completed by a joint mechanical and electrical engineering team that were able to use the project as their senior year capstone project.
As team leader I was initially focussed on getting the Software, Electrical, and Mechanical design teams started on the design process. I laid out a rough schedule that would result in the vehicle being ready for initial testing 3 months prior to the competition. This would give ample time for testing, any necessary redesign or optimization, and practice of competition tasks. As this was a university project, I wanted to make sure that our team members not only developed an effective design, but could maximize their own learning while doing so. I encouraged the team to develop their systems from scratch, as long as they could identify an acceptable off-the-shelf replacement component that could be purchased in the event the custom component failed or fell behind schedule. In turn, I would insure the funds would be available to do so. This strategy kept the project very engaging as team members had the freedom to design, build, and test systems of their own design, greatly increasing the educational value of the project.
Once the team was on their way, I set my focus on acquiring sponsorship. Based on input from the lead engineers in the Software, Electrical, and Mechanical design teams I assembled an estimated budget of roughly $26,000. I approached the companies whose products we intended to use that year, and explained our intentions, past experience and performance, and what we could provide them in return (publicity and experience with their product). The venture was incredibly successful, and all $26,000 was raised by the end of December. The extra effort that was placed on generating an accurate estimate of the team's anticipated expenditures allowed us to purchase the necessary resources to complete design and construction of the vehicle by the self-imposed deadline set at the beginning of the year.
The team took 3rd place in 2011. The rover performed very well, and was extremely reliable; the judges and other teams were very impressed with the design. The limiting factor in our performance was primarily our video quality, which made it difficult to distinguish fine details in the returned images, often costing us valuable competition time. This limitation made the equipment servicing task very difficult, despite a great deal of practice. 2011 was the first year the team experimented with digital video in the hope of getting a less noisy signal, however, due to the higher frequency, the signal was more easily obscured by terrain and forced the team to drive solely on GPS for a portion of one of the tasks (in itself an impressive feat). The sample return task also posed some problems as the sample analysis was approached incorrectly.
Below are pictures of the 2011 OSU Mars Rover showing the arm, camera, and cargo rack both in Utah (left), and Oregon (right).
2011 Oregon State University Mars Rover Design Report.pdf