ROUGH TERRAIN ROBOT SUSPENSION AND DRIVE
June 2017 - May 2018, for MRover. Suspension and drive for use in desert terrain.
Problem and Background
The University Rover Challenge is very fast-paced; the rover may have to cross 1km of desert terrain in 30 minutes while completing complex objectives. For URC 2017 and before, MRover built a six-wheel rocker-bogie suspension and drive system similar to that used on all of NASA's Mars Rovers. However, while this design is great for stability and redundancy, it is not well-suited for speed due to the poor handling of dynamic loads in the rigid links of the rocker-bogie linkage.
For URC 2018, I decided to prioritize rover traversal speed in the pursuit of more points at competition.
Solution
In order to enable higher traversal speed, I implemented non-rigid suspension in the rover's mobility system while moving from six to four wheels to stay within the mass constraint.
Design Goals
- Minimize mass
- Provision for individual wheel yaw control (wheel steering)
- Maintain ground contact for each wheel when one is 40 cm higher than the plane formed by the others
- Maintain a 30 cm ground clearance from the bottom of the chassis of the rover
- Allow the rover to safely traverse typical Utah desert terrain at 2 m/s
Reflection
This system is an example of a bad idea executed well. I often use it as an example piece in training younger members of the team in how easy it is to lose track of the requirements. See the following details to understand.
Successes:
- Efficient manufacturing due to extensive use of waterjet parts (quick turnaround)
- Repair and servicing was simple due to standardized fasteners and abundance of spare waterjet parts
- The system worked well at competition (the team earned 9th place -- our first time in the top 10)
Failures:
- The suspension system was only well suited for "drop case" loads, or forces from directly below the rover. In reality, the competition environment presents many instances of what we call the "curb case", or loads coming from the front of the rover while it tries to surmount large rocks. These loads did not fall in the "plane of compliance" of the suspension system, or the plane that the suspension linkage moves in.
- The needle-roller thrust bearings were exposed to the environment and became ineffective due to lodged sand
- The custom shock was not manufactured to tight enough tolerances and would easily bind and leak oil. In application, we filled the shock chambers with foam instead of oil and abandoned sliding fluid seals.
Response:
The following year, the team redesigned the suspension system to still use four wheels with independent suspension on rockers, but reoriented the plane of compliance to include frontal loads.