Fleet Safety – Looking to 2022 trends in ADAS and Telematics


Sharing is caring!

The National Safety Council in the UK reported that motor vehicle deaths are estimated to be the highest in 13 years, despite dramatic drops in miles driven due to the pandemic. The preliminary estimated rate of death on the roads in 2020 spiked 24% despite miles driven dropping 13%. With these numbers in what was an ‘off year’ for driving, the fleet industry faces a mounting risk problem.

Yoav Banin, Chief Product Officer at Nauto and an expert in fleet safety technology, predicts that in 2022, the adoption of video based vehicle safety technology will be at its highest level. Traditionally, video based safety meant recording and sending “after the fact” reports to safety managers such as when a driver hard brakes.

Yoav believes that the major trend going forward is predictive and preventive AI that alerts drivers in real time, in advance, to avoid collisions versus relying on post-event reporting after an incident has occurred. The most successful AI will limit false positives and reliably alert drivers in advance with enough critical extra seconds to reduce the risk of a potential catastrophic event. Yoav shares some his 2022 predictions on why this type of preventive AI will be imperative for the coming year and beyond:

  • Deeper Understanding of Risk – For example, ADAS (advanced driver assist systems) typically have forward collision warning capability and can alert a driver if they are distracted. Depending on the scenario, such as a driver that is distracted for a few seconds on an empty, rural road – the risk levels go down. However, if a driver is distracted witha a combination of factors that are influenced by variables such as speeding, entering a crowded urban area with pedestrians, or harsh weather – the risk of an accident increases exponentially. The next layer of intelligence can determine what is important and not important – for collision avoidance and for an improved driver experience — by fusing different risk factors, enriching context beyond basic telematics detection and single isolated events.
  • Next Level Operational Point of View – Fleet safety technology is generating tons of event data without context to the cloud and is overwhelming fleet managers. This translates to higher operational cost as companies try to scale the technology across their organization by requiring more time and personnel to sift through and manage the data to discern which drivers need coaching and in what capacity, e.g. what specific behaviors need to be addressed based on the data? Addressing true risk factors on a per driver basis, as opposed to a generic bucket of risk, will drive bottom line results.
  • Predictive Accident Prevention – Understand risk factors that are not tied to an imminent collision – such as progressive drowsiness, where the driver isn’t a risk yet, but could become so over a long haul delivery route, is a preventive, second line of defense – almost a co-pilot – for more advanced AI technology. In real terms, a supervisor may receive a series of warnings about a particular driver enabling them to check in, have a live, over-the-air conversation and possibly advise taking a break.


  • Rise of the Driver – The adoption rate among drivers of this preventive AI technology will rise in 2022. On a macroeconomic level, the demand for more drivers means onboarding less experienced drivers quickly to address the current staffing shortages and supply chain woes. Fleets will need to adopt safety technology that is driver-centric, meaning it is imperative that a driver feels supported and protected by the technology and not monitored 24×7. Striking the right balance between driver privacy and risk is key.
  • Leveraging Data for Traffic Flow – Municipalities want access to shared data that can create maps and build location based risk models to create alternate routes and traffic configurations that decrease accident prone hotspots. This real-world data and driver content can also have implications on better insurance underwriting models.