Techniques are described herein for precisely determining conditions of accident patients, including assessing severity, determining a best trauma center, and deciding the best route to take to the trauma center. Vehicle-to-Infrastructure (V2I) contextual emergency event classification, scoring, and response may be achieved by obtaining speech or video input provided by an Emergency Medical Services (EMS) operator and running the input through a Recurrent Neural Network (RNN) (for speech input) and/or a Convolutional Neural Network (CNN) (for video input) to convert the input into keywords. The keywords can be run through a Multilayer Perceptron (MLP) to identify certain medical-specific keywords to determine a category of doctor and a severity (e.g., Level 1 to Level 6, where Level 6 refers to least severe and Level 1 refers to most severe). The category of doctor and severity are then input into a hospital determination algorithm along with time and location data. Once the appropriate hospital is determined, that data is fed into a hospital routing algorithm to route the ambulance to the hospital. Unlike standard Global Positioning System (GPS), this routing procedure also uses road design data, which includes additional information regarding road conditions (e.g., whether the road has speed bumps, whether the road has a physical divider, etc.).
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Sane, Purvesh and Dahir, Hazim, "VEHICLE-TO-INFRASTRUCTURE CONTEXTUAL EMERGENCY EVENT CLASSIFICATION, SCORING, AND RESPONSE", Technical Disclosure Commons, (March 31, 2020)