PRISM Updates: Lessons learned from these previous attempts

Speaker: Richard Stansbury, student, EECS, University of Kansas, 2002.

Modified Transcript: Dante I was primarily a walking robot which was good for its purpose of scaling Mount Erebus. One of the key things we learned from the Dante I work is the advantage of using sensors, such as laser range finders, which can quickly scan an area. This is especially beneficial if you have some terrain that you think you have a good idea about, but have not traversed before. These sensors can really help you know what is in front of you and thus are crucial for making good remote decisions about what the robot should do, or where it should go, next.

Nomad taught us that certain sensors, like laser range finders, work pretty well in most cases. It also showed us that sensor failures are common. For this reason, we are going to need to have a lot of redundancy in our robot. Nomad also proved that some of the sensors we suspected might not work in this environment, won't work. An example of this is using a camera as a navigation instrument. Cameras are pretty much useless from a navigation point of view, because everything in the environment is white. Cameras will only be of use for sending pictures back for outreach purposes.