Robot-human collaboration has gained ground in various applications due to the flexibility it provides when it comes to hazardous situations. During emergencies, the environment is altered, which makes it tough for humans to traverse; however, through robot-human collaboration, this can be circumvented. The use of human-robot collaboration in underground mine environments during underground mine emergencies could reduce the fatalities associated with underground mine emergencies. With this technology, robots can assess the conditions underground and help detect miners underground, thus preventing human rescuers from entering such hazardous zones.
The Mine Sustainability Modeling Group, with funding from the Centers for Disease Control’s National Institute for Occupational Safety and Health (NIOSH) is conducting research to facilitate human-robot collaboration for miner self-escape. This research effort includes developing algorithms for autonomous robot navigation underground and object detection capabilities for underground robots.
Robot-human collaboration requires some form of communication between the robot and the human. Visual communication is one of the key requirements in such teaming for rescue operations. For this to be done intelligently, there is the need for object detection algorithms, although these algorithms require area-specific datasets to be trained on for their efficacy.
However, due to the safety-conscious nature of underground mine operations, it is challenging for researchers to access such mines and generate the necessary datasets for developing object detection algorithms. The visual perception of robots during emergencies is a key component, and the unique dataset developed by the Mine Sustainability Modeling research group at the Mining and Explosives Engineering Department at Missouri S&T plays a crucial role in this. This dataset captures postures in underground mine environments in various scenarios, including working, resting, and emergency situations. Figure 1 shows sample of the collected images.
Figure 1 Sample images
This data set is available for researchers to use for further development. If you are interested in using the data for further research and development, fill out this form to request access to the data.