Automotive telematics provider VisionTrack has created an IoT video telemetry platform using a MongoDB non-relational (NoSQL) database.
The company’s cloud-based, device-agnostic solution, known as VT2000, is designed for companies managing large fleets. The aim is to give them near real-time monitoring of vehicles and driver behaviour to improve safety.
It relies on MongoDB’s database to store the large volumes of unstructured data produced by a vehicle’s on-board connected camera, such as videos and snapshot images.
Immediate access to driver data
VisonTrack executives claim VT2000 can not only improve driver safety but also significantly increase the speed at which insurance claims are resolved, by giving fleet managers immediate access to driving data, such as speed, steering and braking, as well as images and video from any incidents that occur.
The system sends telematics data from each vehicle to the Microsoft cloud-based platform every 15 seconds for storage and retrieval.
Should a driver be involved in an accident, fleet managers and First Notification of Loss (FNOL) teams are alerted, giving them the opportunity to assess the damage and alert the appropriate emergency services.
Related: The IoT needs a new kind of database
The immediate notification function has already benefited one customer, VisionTrack claims.
The company said an insurer had reported a zero-percent disputed claims ratio and was able to conclude all fault claims settlements within 24 hours – down from three to four months – after using VT2000.
In a company statement, Simon Marsh, VisionTrack’s managing director said: “Having clear real-time data on vehicles is obviously hugely beneficial. Due to the accuracy and quality of the software, very few people will question the result of a claim when we use this technology.”
“But as good as better insurance processes are, what my team really focuses on is modifying driver behavior, improving safety and, ultimately, saving lives. It’s all about prevention and we’re seeing drastically different driving behaviors when these systems are installed.”
“If individuals understand their driving is being monitored, they think more about their actions. When our devices pick up speeding or harsh braking or acceleration, they also give an audible alert and snapshots are sent to the cloud to be reviewed. Such snapshots help to stop unneeded conflict with a driver because they give a clear view of what happened,” Marsh said.
VisionTrack told Internet of Business that companies using the VT2000 include private insurers, such as Cubit Insurance, Markerstudy Group and Signature Underwriting.
Relying on MongoDB
To get its system up and running, VisionTrack implemented the NoSQL database provided by MongoDB.
One of the main benefits of IoT is the opportunity it gives companies to store and analyse real-time data to extract useful insights. In an application like VisionTrack’s IoT video telematics platform, having this kind of data is crucial, which is why a flexible and scalable database is important.
For this type of IoT application, the advantage of using a flexible NoSQL database like MongoDB’s over a rigid relational (SQL) database, typically provided by the likes of Cisco or Oracle, is that it can cope with managing the volume of heterogeneous data generated by numerous IoT devices.
As opposed to a SQL database, NoSQL databases do not require predefined, fixed schemas to collect mixed data types, meaning it was easy for VisionTrack to add new features and data types during the planning phase.
Specifically, VisionTrack is using MongoDB Enterprise Advanced to ensure its data is secure and compliant, and OpsManager to simplify the operations of running the application’s data layer.
In comments to IoB about the benefits of the solution, Karl Wootten, solution architect at VisionTrack, said: “The first big advantage, in using MongoDB over a relational database was the flexible schema.”
“Having no pre-determined data structure meant we could incorporate any type of data into the platform – be that geo-location, video, images or time-series data like the speed of the vehicle. Not only can we process that wide variety of data but, crucially, we can get it in and out of the system quickly. That’s something we would may have been unable to do quickly and flexibly using a traditional relational database.”
Related: Vehicle recognition technology drives targeted roadside advertising
Video data ‘doesn’t lie’
VisionTrack told IoB that the companies started development in March, and the solution began monitoring vehicles in September, meaning it took six months from development to the initial go-live. And the team is continuing to build out functionality in the platform.
The company confirmed it is working on developing a system for consumers for a Q2 rollout, but for now the main focus is on helping businesses.
VisonTrack’s Marsh concluded: “Whether you’re a fleet manager or a parent, you can see how a driver drives in the real world through such technology and this can help stop any ‘wannabe’ racing drivers before they cause an incident.”
“Statistical data, supported by video, doesn’t lie – and this forms the perfect system to help protect all drivers and encourage better driving standards on our roads. Thousands of people are seriously injured every year, and I believe a combination of telematics with integrated video and good road safety education can play a big part in reducing these numbers and preventing incidents from occurring.”
Related: Couchbase & Verizon: why the IoT needs ‘schema-freedom’