Hybrid Site Sensing and Human-Multi-Robot Team Collaboration for Disaster Relief at Nuclear Power Plants

Abstract

Robotic technology is important for creating spatial maps of harsh environments such as disaster sites. A virtual scanning environment is created by using a 3D CAD (Computer Aided Design) model of a nuclear power plant site and simulating various disaster scenarios such as reactor explosions and earthquakes using a physics engine. A virtual mobile robot is used to move around the disaster site and collect laser scans by carrying out raytracing on a voxel grid. This research also introduced an online segmentation method using Multi-View Context Pooling (MCP) for mobile robots to incrementally build a semantically rich 3D point cloud of the environment. During the post-processing stage for the acquired point cloud data, a three-step approach of segmentation, classification, and merging for object recognition is applied, such that each object is precisely represented as a cluster of points that have a unique class label.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 31, 2022
Accession Number
AD1174962

Entities

People

  • Jun Ueda
  • Yong K Cho

Organizations

  • Georgia Tech Research Corporation

Tags

Communities of Interest

  • Autonomy
  • Cyber
  • Space

DTIC Thesaurus Topics

  • Autonomous Navigation
  • Civil Engineering
  • Computational Science
  • Computations
  • Computer Vision
  • Computer-Aided Design
  • Construction
  • Control Systems
  • Corporations
  • Detectors
  • Disasters
  • Engineering
  • Nuclear Power Plants
  • Object Recognition
  • Pattern Recognition
  • Point Clouds
  • Three Dimensional

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Computer Vision.
  • Emergency Management and Homeland Security.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - Bayesian Inference
  • Autonomy
  • Directed Energy