Abilotti: Created page with "'''Shock events are triggered by using real time accelerometer data patterns''' The analysis is made directly on device and does not require any 3G/4G connectivity or for the..."

2020-07-07T22:54:48Z

Created page with "'''Shock events are triggered by using real time accelerometer data patterns''' The analysis is made directly on device and does not require any 3G/4G connectivity or for the..."

New page

'''Shock events are triggered by using real time accelerometer data patterns'''

The analysis is made directly on device and does not require any 3G/4G connectivity or for the system to have GPS coverage.

'''There are 3 distinct types of Shock Triggers:'''

:*Hard Acceleration '''(HA)''' - triggered by a very abrupt acceleration
:*Hard Braking '''(HB)''' - triggered by a harsh braking
:*Hard Cornering '''(HC)''' - triggered when the vehicle takes a sharp curve with high speed

'''Note: In order for the Shock Triggering mechanism to work properly, a vehicle should by correctly assigned to its proper vehicle class'''

'''Vehicle Classes are:'''

::#Very Light Vehicle (under 2000 pounds)
::#Compact Vehicle (2000 - 3000 pounds)
::#Midsize Vehicle (3000 - 3500 pounds)
::#Large Vehicle (3500 - 4500 pounds)
::#Compact Truck & SUV (4500 - 5500 pounds)
::#Midsize Truck (5500 - 7000 pounds)
::#Large Truck (7000 - 8500 pounds)
::#Very Heavy Vehicle (more than 8500 - 10000 pounds)
::#Extremely Heavy Vehicles (Used strictly for Fire Trucks)

After assigning a vehicle with it's proper vehicle weight class;...
* if the client has too many or too few shocks events, sensitivity can be adjusted by the client or by Idrive Support Team in the iris platform (see image below):

::::::::::::'''iris – Fleet Manager – Vehicles – Configure'''

[[File:FM.vehicles.config.png|750px]]

'''Notes:'''

*If the client requires adjustments that are out of the pre-established parameters for sensitivity, please inform Idrive Support Team and they will communicate this need to Idrive Engineering.
*When a vehicle is in sleep mode, the accelerometer settings are more sensitive so the camera can catch an unoccupied hit-and-run or other G-Force abnormalities.

'''Accident Prediction'''

Every shock triggered by the AI Camera is '''processed by a "Neural Network"''' which estimates the probability of that Shock to be an "accident"

The algorithm is trained to "pattern recognize" accidents vs normal shock events

If the probability is recognized to be high the event will be "marked and appear" in iris as a '''"Possible Accident"''' Event [[File:Iris events.png|center]]

<br/><br/>

Shock Event / G-Force - Revision history

Abilotti: Created page with "'''Shock events are triggered by using real time accelerometer data patterns''' The analysis is made directly on device and does not require any 3G/4G connectivity or for the..."