Autonomous guidance of planetary rovers
Operating a rover on another planet is challenging. Gravity might be quite low, reducing the traction that wheels can generate, while dust traps, unexpected ice deposits, and rough or sloped terrain can all be expected to occur.
Nonetheless, the mission of most rovers will be to find and approach sites of scientific interest provided that the goal is judged to be worth the combined expenditure of time, energy, opportunity and risk in getting there.
On Earth, this is a question that can be tackled quite easily as we rapidly evaluate the advantages and disadvantages of following competing paths, but on another planet the situation is quite different. If our rover has to report its situation to Earth every step of the way, and then wait for instructions on how to proceed, the time delay in communications greatly hampers progress.
We seek to provide our rovers with the capability to make basic decisions about path planning in situ, and therefore avoid these considerable delays.
We propose to use a technique called inverse simulation to achieve this objective. The rover will be provided, or even identify, a distant target, and it will then use an algorithm to determine the control inputs, such as acceleration and steering commands, that will bring it to that objective along the most direct, fastest or safest route.
The chosen path can even represent a compromise between these drivers, depending on the weighting applied to each consideration along the route.
Why is this research important?
Developing an algorithm that accounts for both the kinematics of the rover, so that the effects of different commands may be predicted, and which can weigh the importance of the factors governing the desired path, is a complex mathematical task.
At the same time, the algorithm must be flexible enough to identify and cope with unexpected situations in an appropriate manner. However, this approach – coupled with new visual systems – will greatly enhance the capabilities of our planetary rovers in the future, allowing them to cover much greater distances, identify more interesting scientific targets, and ultimately help to provide us with the information we will need to land astronauts on planets beyond the Earth.