Telematics - ML-based virtual sensor for Vehicle Load detection

Our customer, one of the leading Telematics data aggregators for commercial vehicles, wants us to improve their vehicle performance monitoring by including the load status of the vehicle to increase the operational efficiency

Challenges Faced:

● CAN feed data does not contain any direct load sensor information

● Vehicle load status has to be derived from existing CAN feed data points

Decision intelligence

● Machine Learning System for automatic load detection using RPM, Torque and Gear ratio values, acting as a virtual sensor

● K-Means clustering algorithm for Load, No-Load detection

● Adds Load dimension to the vehicle performance monitoring system

Outcome delivered

● Our customers (OEMs and Fleet Owners) could monitor vehicle performance through ML-based virtual sensors, even in the absence of physical sensors

● Zero hardware upgrade cost across the fleet for load detection and enhanced monitoring facilitated through existing infrastructure