Indian Company’s AI Tech Cuts Accidents By 50%

Netradyne’s AI-powered device, developed by Avneesh Agrawal and David Julian, has been reducing road accidents among commercial fleets across the world. With the help of advanced AI and sensor technology, the system tracks driver behaviour and environmental conditions, offering real-time alerts to ensure vehicle and occupant safety. Many of Netradyne’s global clients have reported accident reductions of up to 50%.

The fact that accidents are being reduced by half in a matter of months might sound unbelievable, but Avneesh Agrawal claims that this is precisely the kind of change their AI solution is bringing to fleets worldwide. In countries like India, known for having some of the highest road fatality rates globally, the potential impact could be huge.

Avneesh, who previously spent many years at Qualcomm, including as head of research, holds over 200 patents. Together with his Qualcomm colleague David Julian, they co-founded Netradyne in 2015. The company now employs around 800 people globally, with a significant portion of its R&D team based in India. It serves over 3,000 clients, all of which operate commercial fleets—ranging from taxis and school buses to long-haul interstate vehicles. Their customer base span across the US, India, Canada, the UK, Australia, New Zealand, and Germany.

The safety solution is equipped with a variety of sensors, including GPS, LTE connectivity, multiple cameras, and an inertial measurement unit- bound together by software that runs complex algorithms at the edge of the system. According to Avneesh, this edge processing allows the AI to understand the environment in real-time.

How Does Netradyne Work?

The AI system can detect when a driver is distracted, drowsy, on the phone, tailgating, or ignoring traffic signals. The device then issues alerts through audio cues like a beep and simultaneously notifies fleet safety managers. Over time, the system tracks a driver’s behaviour and assigns a “green zone score,” indicating the likelihood of that driver being involved in an accident.

The latest iteration of Netradyne’s system also includes micro-weather detection. It can sense snow on the road, foggy conditions, and more. By combining driver behaviour data with weather conditions, the AI machine adjusts its alerts accordingly. To date, Netradyne’s technology has logged over 14 billion miles, currently tracking about 600 million miles each month. With such a vast amount of data, including thousands of accidents, Avneesh says they can now accurately forecast accident risks for drivers and entire fleets over the next year.

Beyond immediate alerts, Netradyne’s system also encourages long-term improvements in driver behaviour. It provides video snippets highlighting risky behaviours that drivers can work on to reduce potential hazards. With fleet safety managers also keeping track of, drivers are incentivized to stay cautious. According to Avneesh, some clients have seen a 90% drop in distracted driving-related to mobile phone use within just three to four months of adopting the system, and many large fleet operators have reported a 40-50% reduction in overall accidents.

Looking to the future, Netradyne is evolving its technology towards generative AI. Until now, their systems relied on traditional AI techniques like deep learning, which involves training models on vast datasets of specific scenarios. Now, Avneesh and his team are building a generative AI platform leveraging the massive amounts of data they have collected.

This new platform aims to predict the next 2-3 seconds of driving scenarios with high accuracy. Though still in its early stages, Avneesh notes that the results are promising. Generative AI models are beginning to exhibit behaviour that mimics human reasoning, enabling them to anticipate situations they were never explicitly trained to handle. This advancement, he believes, will take Netradyne’s safety technology to the next level, offering even deeper insights and guidance for drivers.