CrowdHelp: M-Health application for emergency response improvement through crowdsourced and sensor-detected information

Abstract

Preventing natural disasters is beyond our capabilities, but providing better information to disaster management professionals (DMPs) and affected persons is not. Current disaster management systems get their primary inputs from 911 calls and from observations of first responders. Typically such interactions do not follow a prescribed scenario and they do not produce uniform results. Additionally, this process is slow and cumbersome and subject to transcription error [2][12]. We propose an expanded information gathering and distribution tool which uses crowdsourcing to deliver more accurate information to disaster managers more quickly than can be done with existing systems. Using our system, CrowdHelp, people within the radius of a natural disaster are able to send text, pictures, videos, locations, and descriptions of what they see. Our software analyzes the data received, authenticates the sender, removes inputs that are likely to be malicious, clusters reports by type, urgency, or location as desired by the human operator, then displays the results on a map along with suggestions to the operator concerning what type of help is most needed. CrowdHelp also collects additional sensor information from smartphones for future analysis by professional disaster management organizations.