Robotic Perception for Underground Rescue

Robots are potential tools for life saving in underground rescue
operations like mine disasters. Human rescuers are thwarted by roof
falls, explosion dangers, quality of air, visibility through smoke and
dust, mental stress and physical endurance. The inability of human
rescuers to cover sufficient distance in a short time often has fatal
consequences for accident victims. Robots which autonomously scout
ahead of human rescuers and summarize environmental information can
greatly enhance situational awareness, enabling teams to push forward
without delay.

This project envisions immersive, robotically-created 3D models, fused
from many sensor sources and created through, smoke, dust, mud, flood
and fire. These expansive models can provide rescuers with visual
clarity that is uncorrupted by environmental condition. Novel
perception, incorporating radical new modalities of LIDAR, RADAR,
multispectral imaging and actively-illuminated RGB sensing will be
explored. Robots using these approaches will build and analyze models
to detect humans and environmental dangers.