Patent attributes
A system for generating emergency egress advisement based on motion data and trained models for the building. A computing device can receive, from first detection devices, frequency of movement indicating routes frequented the most by users in the building, determine an initial emergency egress advisement for the building, receive, from the first detection devices, first motion data about the users, receive, from second detection devices positioned throughout the building or worn by the users, second motion data, iteratively train a model for the building by correlating the initial emergency egress advisement to corresponding (i) first motion data, (ii) second motion data, and (iii) historic user presence information across one or more model layers using one or more machine learning algorithms, the iterative training generating a trained model for the building, and update, based on the trained model, the initial emergency egress advisement for use during an emergency in the building.
