ESA title


  • ACTIVITYDemonstration Project
  • STATUSCompleted
  • THEMATIC AREATransport & Logistics, Finance, Investment & Insurance

Objectives of the service

Figure 1 Dashboard showing typical data to form the AFNOL User Interface.

Within an hour, vehicle insurance companies would like accurate electronic data relating to a crash or incident so they can decide how to handle the notification of the event internally and to third parties like vehicle recovery and the ambulance service. Currently insurers are dependent on phone calls from the insured driver and possibly an alert from a tracker that an incident has occurred. The result is that the insurance company lose control over the recovery and repair of the vehicle.

AGM Technologies provide high quality data to the insurance company from a battery powered windscreen mounted device (Fig2). The aim of the project is to provide the improved analytics to maximise the use of the data (Fig 2) to give as complete picture of the accident as possible to the insurance agent.

Users and their needs

The users of the AFNOL service are the claims handlers and data analysts within vehicle insurance companies. The data analysts need accurate and sufficiently granular data to start to apply Artificial Intelligence/Machine Learning so they can begin to understand the meaning of the data. For example, is it a real crash or not and the severity of the crash with regards to potential injury? Using an AI enhanced view they are looking to the analysis to be able to settle claims quickly and confidently as possible.

The challenges for the project are

  • to ensure the data sent to the insurance company is pre-qualified by removing misleading or false positive data.
  • to provide the impact analysis showing severity and angle of impact. Deduce the likelihood of injury to driver and passengers.
  • pull together all the data needed to determine if a fraud has taken place
  • put in place the data platform from which AI/ML can grow

The need is international with a known demand from South Africa and Asia.

Service/ system concept

The system user is provided with a user interface which will provide clear and simple advice to the user in the event of a crash which allows them to see:

  • Type of crash
  • Severity of crash
  • Potential likelihood of injury
  • Estimated cost and liability of the incident.

From this information the user can manage the internal and external communication of all aspect of the crash.

The overall system takes enhanced positional and timing data from the in-vehicle device iC6/7 (Fig2) and aggregates it to produce the situational awareness required. The system’s enhanced data includes increased positioning sampling frequency from 30s to 1Hz and the accelerometer sensor sampling rate increased from 100Hz to 400Hz. This increased sampling rate or granularity is sufficient to allow the generation of the Machine Learning models.

The software development activity in the project includes

  • changes needed to handle the enhanced sampling rates and the addition of other external data feeds like weather.
  • New user interface as described.

Space Added Value

AGMT make full use of GNSS to provide position and compute speed. A strong relationship with SONY has allowed the delivery of a sampling rate of 1Hz.

Current Status

Figure 2 Data collection device (IC6) to feed AFNOL analysis.

Held detailed workshop with key customer which provided powerful high-level needs for their business. These statements concur with our project proposal and demonstrate that the timing of the project is relevant and timely. Their nomenclature for the future system roll-out is ‘eFNOL’.

Given the global silicon supply shortage, the company has secured the manufacturing continuity of the IC6 data collection device after a re-design using available components.

The project has completed all planned stages and has carried out the pilot with a detailed crash demonstration and analysis. The results of the project are detailed in the PilUP and the FREP. In summary A summary and achievements document has also been prepared. The conclusions areas follows.

The project has substantially achieved the initial objectives albeit with a different pilot to that envisaged due to time constraints. Also in the time available the number of crashes that the pilot companies could gather data on was unknown and likely to be small due to the relative rarity of crash events. Therefore conducting our own multi vehicle live crash scenarios was very effective and efficient.

The technology advances of using a wholly new ‘privacy preserving’ audio detection solution with the added piezo sensor were tested first by using sled testing then with real vehicles in the pilot. The results have been significant in achieving the objective of reducing the false positives by over 99% which would have been in the data monitoring the vehicles behaviour.  The new technology produced additional benefits of being able to monitor incidents to the vehicle like side scrapes which would not normally be seen by a conventional telematics system. Furthermore, the high suppression of false positive data minimises the data transmission by over 95% thus reducing cost and battery life.

The project also investigated the use of ultra-low power analog RAMP technology to continuously process the microphone and piezo inputs. This approach will allow the sensor to be “always on” thus allowing the system to capture events like wing mirror hits even when the car is parked.

The new ‘audio input only’ solution that was developed as part of this project also provides the scope for exciting ‘Smart City’ based applications such as a network of ultra low power sensors deployed at road junctions and other high risk road positions which are tuned to listen for the characteristic sound of vehicle collisions (or even other types of events such as glass breaking / gunshots etc). We investigated this briefly during the project and you can see the effect here with the model running on an Aspinity development board – the green LED flash indicates the detection of the event from the audio.

The exposure to the Insurtech marketplace through attending several trade shows has increased our understanding of the market and increased our prospect list. It has also built up our list of prospective resellers. The results of this project have enabled AGMT to be featured in the Insurtech ITC DIA Community top 100 companies to watch. The entry will be found in the full list at

As a result of the breakthrough on crash detection and faster sampling rate we now have a highly competitive service offering improved First Notification of Loss service offering. We can also offer new services to customers including; proof of liability for ‘hit while parked’ claims and red/green return lane for rental cars.

Prime Contractor(s)

Status Date

Updated: 14 February 2024