Autonomy Community

AI-based Damage Detection a Game Changer for Fleet Operations

Published on March 2, 2023 by Daniel O'Brien

Undetected vehicle damages are one of the major cost drivers for the mobility service operators. INVERS now adds an AI-based damage detection to its digital shared mobility solutions.

Why Damage Detection?

Only one in eight carsharing customers who cause damage to a sharing vehicle report it to the operator. As a result of this daily occurrence, carsharing operators lose up to 10% of revenue because they do not have sufficient information to claim damages. Enabling them to realize this significant share of revenue by means of damage detection lets them be more profitable and use these profits to increase investment in their services and spread shared mobility even further. Also, damages that are not reported mean that the next customer faces safety risks if their car's roadworthiness is impaired. In addition, operators lose out on the resale value of their cars since they usually do not carry out regular checks, and so do not detect minor and medium damage until it's time to sell their vehicles.


Damage detection is also a challenge for traditional car rental companies who carry out regular, resource-intensive visual inspections that are unnecessary in approximately 90 percent of vehicle returns and usually only detect externally visible damage. So, damage detection and successful damage management is a key challenge for fleet operators in car sharing and car rental significantly impacting profitability. Addressing this issue with a precise, fully digital small damage detection solution can provide fleet operators with a positive business case immediately. 


How Does Damage Detection Work?

Aiming at solving this challenge on a technology level, Invers and carValoo have joined forces, combining reliable connectivity, proven sharing technology, highly sensitive sensors and an innovative AI solution. Having invented the first onboard unit for automated vehicle sharing in 1993, Invers offers integrated hardware and software solutions for mobility service providers to launch, operate and scale their offering. Together with carValoo, a leading provider of AI-powered software solutions for damage detection, Invers is now further expanding its proven solution for shared mobility with an AI-based damage detection. The new solution offers digitized and automated damage detection based on data from highly sensitive motion sensors within the car. This data is analyzed and processed by artificial intelligence to detect damage precisely and reliably. As a consequence, it immediately contributes to business success as it allows operators to reduce costs by improving damage management, increasing successful damage claims and completely digitizing the rental process. 


Details Behind the Intelligent Sensor-Based Solution

Highly sensitive acceleration and gyro sensors in the new solution capture various data from a vehicle in daily use. The motion sensors generate approximately 72.000 data points per minute per car which sums up to 103.7 billion data points per day per car. However, not all of them are related to damages and are relevant for damage detection. For example, closing a door of a parked car or driving through a pothole creates sensor data points similar to a parking accident. The solution’s intelligent algorithm already filters and compresses a large part of this raw data already on the sensor box within the car, so that less than ten percent of the data points still need to be transferred to the cloud. The data points that arrive in the cloud are then further analyzed in detail there by an AI algorithm which identifies patterns and anomalies. The AI then assigns them in real-time to specific events that, like a parking accident, have actually led to damage to a particular component. At the same time, it recognizes if it was only a door being closed or a pothole being driven over, resulting in no damage and making a report unnecessary.


What’s the secret behind the technology?

The key strength behind AI driven technologies, is its capacity to learn from big data sets and its capacity to constantly improve. Since the application of the solution 3 years ago, damage reports have been compared with actual damage, so that the artificial intelligence continuously adapts and optimizes its accuracy. The system reliably identifies even minor damages and those damages that are not detected in usual visual inspections  in over 95 percent of cases and locates them correctly on the vehicle. This precision in combination with a fully digital solution including sensors, an API and the backend is unique. 


What does it mean for the mobility industry?

With the new solution for damage detection fleet operators gain very precise insights into the condition of their vehicles so they can automate their damage management. AI-based damage detection enables them to detect damages,  assign them, claim for them and increase the efficiency of maintenance and services. It has the potential to become a major game changer for fleet operations and is another example of how the mobility ecosystem can benefit from innovative technology to ultimately make the use of shared mobility more convenient than car ownership.