Farming today is more complex than it has ever been. Operators are increasingly reliant on technology to aid and improve harvest performance. New harvest technology is under development that will advise harvest operators on the proper adjustment of machine harvest settings, as well as automatically adjust these machine settings without operator intervention, improving the harvest performance of the machine, and reducing the cognitive load of the operator. In this work a high-fidelity, interactive harvest combine simulator is used to understand how harvest operators currently use existing harvest technology, and to evaluate the performance improvements provided by new prototype machine control algorithms and human control interface designs. The interactive harvest simulator is used to assess an intermediate advising step for machine controls adjustment compared with a path using fully autonomous machine adjustment. Testing novel harvest technologies using the virtual environment of the combine simulator introduces a specific set of constraints and challenges that are not found in most other vehicle simulation applications, including the need for accurate physical and visual crop flow representations and a requirement for realistic machine responses to a wide variety of operator input commands. Using a high-fidelity combine simulator for testing allows unique harvest scenarios to be repeated by experienced operators in a controlled virtual environment. This study evaluates operator acceptance, performance, and feedback for two novel pieces of harvest technology, Advisor and Director. Advisor is an operator-in-the-loop system providing feedback on proper machine control adjustments during normal harvest operations. Director is designed to continuously monitor the overall harvest health and autonomously adjust the combine harvest settings. In this study, operators harvested the same virtual field twice, first using Advisor, and a second time using Director. Operators overwhelmingly perceived both the Advisor and Director systems as optimizing the harvest performance of the combine and recommended both Advisor and Director. The results presented in this work show that both systems improved the perceived harvest performance, although the Advisor was not as highly rated. Participants recommended the automated nature of Director, and both operator feedback and physiological measures indicates that this harvest technology reduced the cognitive load of the operator. This work demonstrates two main points. First, the interactive combine simulator can be used for evaluating novel harvest technology in the lab. Second, that operators can quickly acclimate to automation within the combine and were able to harvest in a more productive manner when using higher levels of automation.
