The application of Tobler's hiking function in data-driven traverse modelling for planetary exploration

Using data collected from the Meili-I analog crewed mission hosted on a remote Scottish island during August 2023, we analyze GNSS traverse tracks to assess walking velocity in relation to terrain slope. A series of data sampling tests to derive models using a generalized form of Tobler’s Hiking Function indicates these models are only applicable to a similar resolution at which they were derived. Deriving walking velocity at 20-second intervals suggests a linear relationship between walking velocity and slope is useful for grid walking algorithms, but longer sampling intervals (>120 seconds) indicate a greater sensitivity to slope, likely recording long-period affects of exhaustion from prolonged ascending/descending of slopes. Findings are constrained by the limitations of environmental variables during the mission, including variable weather conditions and increasing familiarity with terrain. Applying calibrated hiking functions to grid walking algorithms (i.e., Dijkstra’s algorithm) offers time-optimal paths useful for walk-back contingency planning but is unsuited for planning exploration geology traverses. • Tracking of crew during a mission simulation shows how walking velocity changes with slope • Data sampling tests show hiking functions only apply to a resolution similar to their derivation • Sensitivity of velocity to slope increases when averaging over longer straight line distances • Paths from grid walking algorithms are suited to contingency planning rather than exploration