Dartmouth engineering professor Laura Ray and PhD candidate Joshua Cook want the tractor caravans that bring fuel and supplies across Antarctica to the National Science Foundation (NSF)’s South Pole Station to be able to get there and back in half the time.

Currently, at a savings of $6 million over transportation by air, two caravans of eight tractors make three round-trips per season hauling approximately 2,250,000 pounds of fuel and supplies. If half the tractors are autonomous, drivers can do double shifts in the manned vehicles, doubling the traverse’s daily operation time and enabling more trips per season per caravan without adding more drivers.

Collaborating with the US Army’s Cold Regions Research and Engineering Laboratory (CRREL) and tractor manufacturer AGCO, Cook and Ray are developing simulation models to provide insight into the types of algorithms necessary to allow an automated unmanned tractor to follow a manned leader tractor across the uneven terrain of the Ross Ice Shelf shear zone. When traveling across firm, high-traction terrain, unmanned vehicles should be able to follow safely by mimicking the throttle and gear selections of the manned tractors. However, across softer low-traction terrain, a different strategy is required: The unmanned tractor could leverage the lead tractor’s Controller Area Network (CAN) data to estimate the properties of the terrain between the vehicles and compute a safe trajectory:

The travel conditions of relatively low speeds—approximately 10 mph—and virtually no human or vehicular traffic make these traverses an ideal case for fleet automation. It may not be long before a fleet of a few manned and many unmanned smart tractors can facilitate such challenging and costly missions.