Rapid automated damage detection with autonomous robots


Northeastern University

Northeastern University

Pacific northwest economic region


Project Description

This project utilizes small unmanned aerial systems (sUAS) to rapidly assess damaged structures following a natural disaster. Currently, after a disaster, human personnel perform on-site assessment of the damage of a structure, which may limit how much of a structure may be assessed, potentially puts personnel in danger, may be time consuming, and is difficult to document comprehensively. sUAS will use GPS, cameras, and laser scanners to autonomously circumnavigate a structure and automatically detect and identify any damage to critical infrastructure. The project will develop algorithms to automatically detect a variety of forms of damage, including bent or ruptured steel members, or cracked, crushed, and spalled concrete. The damage detection is based on innovative machine learning and heuristic techniques. Information collected by sUAS can be received and assessed remotely to rapidly develop a model of the damaged structure. This method of rapid and remote response allows for on-site personnel to focus on other emergency needs, while off-site personnel can develop a plan for restoration.

Principal Investigators