Crash Insights

Learn from real-time crash detection and context to generate actionable safety insights.

Introduction to Crash Insights

Sentiance Crash Detection uses device sensors to identify vehicle crashes within minutes of impact. This enables faster emergency response and more efficient accident management.

By leveraging advanced ML models, the SDK distinguishes between actual crashes, false positives, and other driving events.

Key Benefits

• Direct access to detailed crash insights

• Easy integration of emergency services into your app

• Potentially life-saving intervention through rapid detection

• Post-crash data analysis and investigation

Supported Vehicles

• Cars

• Motorcycles

Features Overview

1

Crash Detection

Your app can receive near real-time crash notifications. Each detected crash includes data from just before, during, and immediately after the impact.

Available crash data:

• Severity The calculated intensity of the detected crash based on delta-V and pre-impact speed this can be of the following values:

  • Low

  • Medium

  • High

• Confidence Indicates the likelihood that the detected event is a true crash.

• Magnitude (G-force) The peak G-force measured during impact.

• Delta-V The change in velocity detected at impact.

• Speed The user’s speed at the last known GPS location before impact.

2

Crash Forensics

The Crash Forensics provides detailed analysis of detected crash events in Insights Control Tower and provides a complete overview of the crash event and the events before impact including:

• Map view of the crash location

• Route taken before the crash

• Risky driving events captured during the trip

• Additional road and contextual information

3

Crash Forensics – Full Trip Report

Crash Forensics – Full Trip Report offers an on-demand, detailed breakdown of the entire trip in which the crash occurred.

• Available on demand

• Covers trips up to 30 days in the past (can be extended if needed)

• Includes full contextual data leading up to the crash

Important: Both Crash Forensics solutions require personal and sensitive data to be transmitted to the Sentiance backend.

Additional notes

Phone Placement

• For optimal performance on motorcycles, the mobile phone must be securely mounted on the handlebars.

• In cars, a wider range of phone placements is supported, while still ensuring consistent and reliable detection.

ML Model Variants

• Detection insights vary depending on the vehicle type (motorcycle or car). This is due to the use of different machine learning models, each tailored to the unique physical dynamics of the vehicle.