Wednesday, November 7, 2007

Rover Selection Criteria

It took the group a while to turn the general plan of developing a tele-operated vehicle to support crews in a Mars analogue environment into a more detailed set of constraints. We agreed on a set of goals; augment the capabilities of the research crews, with a tool that is flexible and on which we can test different software architectures. The point being on providing standardized interfaces so teams with various backgrounds can develop various functional modules - i.e. a fancy word for adding different science capabilities to the same vehicle. The focus being on using the vehicle to train the future space explorers by having students and teachers devise and use the rover for scientific goals without worrying about the underpinning technical details. Not unlike a relation scientist have with supporting industries today. Rather then designing the rover from scratch we decided the interesting parts are on the interface between the the rover and the science - trading hardware development (which is interesting in itself but better done by people who already have experience in the field) for a platform rich enough to enable development of interesting software and algorithms. Nonetheless, selecting the right hardware in light of deployment in difficult terrain, but also keeping in mind that it needs tobe familiar to students and teachers we devised the following constraints to select the rover.

Top five criteria:

1. within budget
2. tele-operational, upgradeable to autonomous
3. reasonably all-terrain
4. video camera/wifi
5. flexible control unit

Additional specifications:

All terrain should include protection from fine dust and sand and somewhat rain-proof with insulated wires, robust. With one of them we are going into a serious environment, the rover needs to cope with it.

GPS with easy upgradeability. While the first cut is tele-operation, science objectives and future autonomous mode will greatly benefit from a gps unit.

Wifi unless not available, then radio control.

Video camera on a mast, start with one, write software for two, capable of mono or stereo.

Flexible control unit so rover programming is accessible to students and teachers.

Manipulator capability, arm with one degree of rotation

Laser ranger or other ranging hardware in order to keep the rover safe so we are able to give wide access to it yet avoid hazards in the field.

Build or purchase a storage house for the vehicle that can be located away from the hab and has easy open doors that will not need human intervention.

Power management that will allow vehicle to be fully charged for 2-4 hours, go out, come back, and charge itself, possibly shut down while driving to the goal

With this in the back of our minds we started the selection process. At the end, and after comparing side by side more than half a dozen manufacturers that were close to our budget we converged on Senseta's MAX series. The other two very close contenders were Pioneer P2AT and the Segway robotic platform. All of these were good contenders but after weighing primarily the cost and the other factors, we decided on Senseta. It is a good platform, and uses a regular small form factor PC capable of running widely available tools and hardware. If Apple ][ in its days was a great gift for schools, we think this rover could do the same for schools today, yet it is tested in the field and can perform a multitude of science goals at an attractive price. Being a standard PC it is easy to distribute development among teams not having direct access to it as it supports a decent simulation environment in software. Moreover, Senseta is closely affiliated with the robotics group at CMU which had some impressive record of innovation lately, and they were very enthusiastic to provide the initial support to get our group of the ground.


No comments: