PhD Thesis Defense: Joshua T. Cook

Tuesday, May 2, 2017, 2:30–4:30pm

Jackson Conference Room, Cummings Hall

“Dynamics Modeling and Control of Autonomous Tractors for Polar Traverses”

Abstract

This thesis derives simulation models and proposes three control modes to handle various operating conditions for unmanned, heavy duty tracked vehicles pulling large payloads. The specific application of interest is for the South Pole Traverse (SPT) which provides logistics support for research across the Antarctic continent. The most significant effort is support for the South Pole Station. In their current configuration, SPT uses eight tractors and an eight man crew for each round trip journey from McMurdo Station to South Pole Station. Here, tractors pull sleds carrying ~3 times the nominal vehicle mass. These large loads can create mobility challenges for tractors but are also necessary to provide the desired economic yield for operation. The aim of this thesis is to equip unmanned tractors with a control architecture to handle the current load demands while maintaining timely and efficient traverse operations. If implemented, the estimated savings amounts to ~$2M that can reappropriated.

The beginning chapters of this thesis derive three different models for tractor simulation. The first is a simplified single body model that incorporates the coupling between the power-train and the terrain-vehicle interaction so that loading effects on vehicle operation are present in numerical experiments. This model is validated against experimental data collected in Antarctica. The second derives a more comprehensive multi-body model of SPT's current towing configuration so that payload stability can be evaluated under closed-loop control of unmanned tractors. The third model is a hybrid, multi-body model that equips the tractor with a hydraulic towing winch and captures actuator limitations via hydraulic component models.

Later chapters use these models to explore and develop proposed architectures for a leader follower control mode, a traction control mode, and a winch control mode. The leader follower control mode is tested and developed using the second, multi-body model where payload stability can be evaluated under closed loop control. The traction control mode and winch control mode use a simplified version of the single body model for filtering and the multi-body winch model for controller testing. The effectiveness of all three control modes for tractor autonomy and mobility are demonstrated via simulation.

Thesis Committee

For more information, contact Daryl Laware at daryl.a.laware@dartmouth.edu.