Degree Type


Date of Award


Degree Name

Doctor of Philosophy


Industrial and Manufacturing Systems Engineering


Industrial and Manufacturing Systems Engineering

First Advisor

Michael C. Dorneich


This research investigates how machine operator expertise, strategies, and decision-making can be integrated into operator models that simulate authentic human behavior in construction machine operations. Physical prototype tests of construction machines require significant time and cost. However, computer-based simulation is often limited by the fidelity in which human operators are modeled. A greater understanding of how highly skilled operators obtain high machine performance and productivity can inform machine development and advance construction automation technology. The initial effort of this work was to develop a virtual operator model (VOM) through a combination of human factors and dynamic modeling techniques. Operator interviews were conducted to build a framework of tasks, strategies, and cues commonly used while controlling an excavator through repeating work cycles. A closed loop simulation demonstrated that an operator model could simulate the trenching work cycle. Once a VOM has been developed that is capable of closing the loop to simulate equipment operation, machine assessment can be performed earlier in the development process without physical prototyping, which reduces cost and development cycles. Advancing the state of the art in operator modeling requires models that can adapt and learn. This work investigated approaches to enable a generic virtual operator model to adapt to machines with different dimensions and capabilities without need to tune the model, adapt to changes in the environment based on the operator’s actions, and adapt to differences in operator skill levels. Finally, learning capabilities and strategy models are going to be developed for the VOM, which will enable virtual operator model to understand machine models, learn during operation and choose appropriate strategies for operation.

Copyright Owner

Yu Du



File Format


File Size

163 pages