Browse Prior Art Database

Predictive Crowd sourced Event Profiles for Public Safety Disclosure Number: IPCOM000239775D
Publication Date: 2014-Dec-01
Document File: 5 page(s) / 201K

Publishing Venue

The Prior Art Database


A system and method for constructing predictive crowd sourced public safety event profiles is disclosed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 31% of the total text.

Page 01 of 5

Predictive Crowd sourced Event Profiles for Public Safety

Disclosed is a system and method for constructing predictive crowd sourced public safety event profiles.

In public environments, situations can often arise where someone needs to be alerted to a danger, but no path of communication exists between a person noticing the danger and the people to alert or to the person creating the hazard. For example, for a truck pulling a trailer with a chain dragging on the highway, nearby drivers have no good way to communicate the problem to the driver or to other drivers in the area. In a simpler example, deer may be grazing along the side of the highway and drivers need to know to watch out.

These situations are complicated because the people noticing the danger often lacking the attention / information / time / capability to fully communicate the danger to others in the area. For example, a driver noticing a chain dragging would not be able to text the other driver (even if they knew the driver's number) because they cannot send or receive text while driving. They may only be able to push an alert button or view an alert on the dashboard.

The disclosed method uses a crowd sourced analytic solution to this problem, where an event profile is gradually built up based on incomplete pieces received from human and sensor data. The partial event is referenced against historical dangers to predict the root cause. The problem is then broadcasted to people who should be alerted.

Existing art in this area includes several components of the overall system, such as alert buttons, broadcasted alerts, and collecting sensor data. The disclosed method aggregates the partial sensor and manual input for hazard identification.

At a high level, a centralized server collects partial input such as hazard location, sensor data or text description, and aggregates the partial input into real-time events. The partial event profiles are compared against known profiles for full predicted identification of the hazard. Based on danger estimates, the server takes actions based on hazard thresholds, for example a text broadcast to nearby people, targeted broadcast to the source of the hazard, or dispatching emergency services.

Key Steps:



the central server. For example, a car's nearby object detector sensor sends its data when the driver presses the alert button.

Central server aggregates all alerts received through all channels into real-time event profiles, with consolidation of alerts based on event location and time.

Users report hazards to a central server using any limited means available. For

example, pushing a simple generic "alert" button on their dashboard and taking no further action, sending a text description (possibly through voice transcription) or simple taking and posting/sending a photo without annotation or comment. Manual alerts with integrated sensors send additional available sensor data to



Page 02 of 5



Users can contact the serv...