A Study on the Construction of a Visual Presentation Method that Can Prevent Cognitive Tunneling in Unmanned Construction

One of the problems with unmanned construction is the lack of visual information, which reduces work efficiency to less than half of that in onboard operation. Therefore, methods to provide visual information using drones and image processing were studied in the past. However, the addition of information causes the operator to fall into cognitive tunneling in which the attention is focused only on a specific image. In this study, we attempted to develop a method that can prevent cognitive tunneling and shift the operator attention to an appropriate image according to the working state of heavy machinery. Cognitive tunneling is caused by low visual momentum (which represents ease in information integration between images) and high visual saliency (which represents ease in attention). Therefore, because visual momentum can be improved by presenting the same landmark in different images, useful landmarks for each work state were included in each image. In addition, because humans tend to pay attention to objects that vibrate in the effective field of view, we presented the image of an external camera in the effective field of view and allowed the image to vibrate when the work state was switched. To confirm the usefulness of the proposed method, an experiment was conducted on an actual hydraulic excavator. Although the proposed method did not improve the work efficiency of some operators, we believed that the proposed interface could direct the eyes of the operator to an appropriate image according to the work state.