A multi-phased evaluation of the Iowa Driving Simulator as a virtual proving ground for construction equipment simulation is presented. In Phase I the Iowa Driving Simulator was evaluated in an “open-loop” mode to assess its capability to simulate a typical maneuver common to wheel loader operation, and its viability as a test platform for human subject evaluation of those maneuvers.
A typical wheel loader truck loading cycle involves numerous directional shifts. Cycle productivity is increased if these shifts are executed at full engine throttle. Jerk and acceleration levels associated with full throttle shifts, however, can cause both operator discomfort and spillage of loaded material. Electronically controlled transmissions have the potential to both minimize directional shift times and material loss while optimizing operator comfort. This optimization will require an understanding of the factors which affect operator comfort during shifts. A study was therefore devised to determine those aspects of the motion generated by a directional shift which affect operator comfort. The Iowa Driving Simulator motion system was used to present operators with a series of acceleration time histories which are representative of various shift strategies. The operators rated the relative comfort of each strategy during paired comparison tests. Limitations of the simulator motion system prevented definitive results from being drawn; however, results did confirm shift comfort criteria previously established by the machine manufacturer. Success of the Phase I effort was sufficient to warrant a more in-depth study.
In Phase II a complete VPG environment for wheel loader operation on the IDS was developed and qualitatively evaluated. This VPG environment included a visual model of a mine pit, developed for Caterpillar, Inc. by engineers at its National Center for Supercomputing Applications office, combined with the immersive motion capability of the Iowa Driving Simulator. A real-time dynamics model of a generic wheel loader along with a menu driven interface to the data set used to simulate a particular wheel loader were developed at Center for Computer Aided Design. This combination of programs allows changes to the design of a loader to be rapidly evaluated within a virtual proving ground environment or off-line at an engineering workstation. The machine model was then combined with an implement/soil interaction model, also developed at Caterpillar’s National Center for Supercomputing Applications office. The resulting machine model can be evaluated either off-line at a workstation or driven in response to operator input within the Iowa Driving Simulator virtual proving ground environment. A comparison of the offline model’s predictions of machine response to swept-sinewave steering input is shown to compare favorably with measured performance of the actual machine.