One of the biggest reasons traditional mapping is not reliable is because it does not account for the variability of terrain such as slope, elevation, soil properties, etc. The Department of Defense knew their current method of predicting the speed at which they could travel would not be accurate without addressing this variability. Sometimes their data would indicate a route was possible, when in fact, they would get stuck. And on the flip side, it might show a route as impenetrable, when actually, it was a viable option.
Objective
Reliably predict the speed at which the Department of Defense can traverse their vehicles across a given terrain using simulations together with soil and elevation information.
Method
RAMDO works in conjunction with the simulation model, allowing engineers to input parameter variabilities. Then it builds a surrogate model to approximate the simulation model, which is much more computationally efficient to run. From this, RAMDO is able to produce an output distribution of results.
Mobility Method Example
RAMDO worked in conjunction with a terramechanics simulation. The variation accounted for was of the input parameters of soil type, slope, and elevation. The resulting output distribution was a series of predictive speed maps at any chosen probability level.
Results
RAMDO accurately shows the predicted speeds obtainable at any given probability level. Results are more accurate because of the inclusion of the variation in inputs. Without RAMDO, probability levels were unknown and likely inaccurate speeds because the simulation did not take into account the variability of inputs. In this example, RAMDO was able to account for
the variability inputs (soil type, slope, elevation) and produce predictive speed maps at any selected probability level. Contrastingly, the deterministic map (without considering the input variations), produced a single map of unknown probability levels and likely inaccurate speed predictions.
