EV ArticlesEV Learning

Understanding AI-based road safety solutions for commercial fleets

Established in 2015 in India, fleet safety and management solutions company Netradyne was recently conferred with the FICCI Road Safety Award. The company has its R&D and manufacturing centre in India. Netradyne says that by adopting its vision-based technology, organizations have achieved a 50% reduction in road accidents and over 90% decrease in distracted driving incidents.

Netradyne has raised $197.5 million globally through key investors such as Reliance, Softbank, Point 72, Microsoft Corp, and Hyundai. The company closed INR 600 cr revenue last year and is expecting to close INR 1000 cr this year, according to a company statement. 

In this interaction with Durgadutt Nedungadi, Senior Vice President – India and International Business at Netradyne, we explore how vision-based systems and Artificial Intelligence are being used to increase road safety in commercial fleet operations.

What specific road safety challenges are faced by fleet vehicles that differ from those faced by personal vehicles?

There are two primary differences between personal vehicles and fleet vehicles. Firstly, ownership; a personal vehicle is typically owned by an individual, and thus the responsibility for its safety, optimization, and upkeep falls squarely on one person. This is not the case with a fleet, where responsibility is shared among multiple parties.

The second difference lies in driving conditions. A personally owned vehicle is typically limited to a specific geographic area, with occasional long-distance trips. However, this is not the case with commercial fleets. Additionally, commercial drivers often spend much longer periods behind the wheel, with drives lasting between 6-12 hours. This extended driving time creates unique significant challenges for fleet drivers and the vehicles they operate.

How are Netradyne’s solutions different from other AI-based dashcams? Are there any India-specific features incorporated for deployment with the Indian fleets?

There are several ways in which we differ:

  • One significant distinction is that we were among the first to introduce a dual-camera system. The external camera focuses on the relative behaviour of your vehicle vis-a-vis other objects on the road, and the internal camera focuses on the driver.

In the West, where there’s a lot more consistency in road construction, we recognize speed signs, stop signs, traffic lights, and U-turns. If there is an infraction, for example, crossing a stop sign, an alert is raised. In India, because we don’t have that consistency yet, we’ve limited ourselves to following distance and collision warning from an external camera perspective. The inward camera checks if the driver is wearing a seatbelt, talking on the phone, or distracted. It also monitors if the driver is drowsy or yawning, among other things.

  • The second difference is precision. We possess a valuable asset in the form of over 10 billion miles of recorded drive time footage, which we use to train our devices. We are able to deliver alerts to our customers with 90% accuracy or higher. Detecting a drowsy driver and confirming drowsiness with 90-95% accuracy is what we strive to meet. Other devices may claim to offer similar features, but if they have a lower accuracy rate of 60-70%, false alerts will bombard the user, causing a loss of trust in the device.

We launched simultaneously in the US and India, recognizing the vast difference in driving conditions between the two regions. In the US, the acceptable following distance would probably be 10 to 15 feet, which is not the case in India. Here, we have a reduced following distance threshold. In India, we recognize two-wheelers and auto-rickshaws, and we are in the process of becoming able to recognize cows, as these are all very important aspects of Indian road conditions.

  • Plus, we use cameras of the highest resolution, which give clear videos even at nighttime.
  • We now have the capability to detect environmental conditions like heavy fog during winter and heavy rain in coastal areas. An alert is raised to the fleet manager, who can identify the vehicles operating in that area and make decisions on whether to change the vehicle speed or stop it in severe conditions.
  • Another example is compound alerts, where a single violation like following distance may not be severe, but when combined with other alerts like looking at the phone, it becomes a high-severity alert. These are some of the ways we differentiate ourselves.

Can you provide a few examples where Netradyne solutions have been deployed?

Our solution has been selected by Amazon for their last-mile delivery vehicles worldwide. This choice was made because these vehicles operate in challenging conditions.

Drivers often work long hours to ensure timely deliveries. Amazon also closely monitors driver behaviour, including seatbelt usage and phone usage. We are currently exploring ways to detect smoking in the cabin as well. Additionally, Writer Safeguard, one of India’s foremost cash management companies, has decided to equip their vans with our device due to their need to transport critical cargo.

Are there any risks attached to increased dependence on Artificial Intelligence in ensuring safe driver behaviour?

Here, I would like to explain the difference between active ADAS and passive ADAS. Currently, we are focusing on passive ADAS, which uses the same algorithms as a normal ADAS vehicle. However, instead of giving the vehicle direct control, the driver makes the decisions. For example, with auto braking, it’s important to test accuracy levels under specific driving conditions. On a US highway, it’s acceptable for the auto brake to stop the vehicle immediately from 50 miles an hour, but in India, the auto braking would need to go down gradually from 80 kilometres an hour to 60 kilometres an hour to slow the vehicle without causing a collision.

These methods help reduce risks but require extensive testing and precision to be effective. This is why autonomous vehicles are taking longer to become a reality. At the end of the day, if you can reduce accident levels to anywhere between, say, 20 to 50% using passive ADAS, then that’s a risk-free way to go.

There are numerous last-mile delivery and hyperlocal fleets running on electric 3Ws and 2Ws in India. Can your solutions be applied in these cases?

We are strongly looking into how we can have some kind of technology for the two-wheeler market; We don’t have one yet. But for the three-wheeler kind of minivans and above, the technology is absolutely valid and viable.

Do insurance companies factor in the integration of tech-based accident prevention systems while charging insurance premiums for fleets?

In countries like the US, liabilities are very high, leading to increased interest in solutions like ours. Indian insurance companies are also looking into whether our device can reduce driving risks significantly. This could result in rewarding the customer with lower premiums or certifying the driver for a lower premium. There are various ways of structuring this, and insurance companies in India are currently exploring these possibilities.

Are there any other road safety solutions in the pipeline that Netradyne intends to roll out in future?

Our main focus is road safety through the use of vision-based technology. We aim to provide a comprehensive solution that includes not only safety but also basic fleet management, fuel management, and vehicle maintenance. This way, our customers won’t have to purchase separate solutions for these needs. We may partner with other companies to achieve this goal, but our focus remains on safety.

As of today, yours is an aftermarket solution. Are you also talking to OEMs to bring it as a standard fitment in some vehicles?

Currently, the majority of our business is aftermarket. However, we are engaged in promising discussions with various Indian original equipment manufacturers (OEMs). It has become clear to OEMs that ADAS, whether in the form of purely passive or a combination of passive and active systems, will soon be essential in vehicles. As a result, we are in advanced talks with multiple Indian OEMs, although we have not signed anyone yet.

Corrigendum: In the August issue of our e-magazine, an article titled ‘Gaining Insight into AI-Based Road Safety Solutions for Commercial Fleets’ included a statement attributing INR 600 crore as Netradyne’s Annual Recurring Revenue. However, this information requires clarification. To be accurate, the figure of INR 600 crore actually represents the company’s revenue for the previous year, rather than its Annual Recurring Revenue.

Also Read: Minus Zero unveils India’s first fully autonomous vehicle

Subscribe & Stay Informed

Subscribe today for free and stay on top of latest developments in EV domain.

One thought on “Understanding AI-based road safety solutions for commercial fleets

Leave a Reply

Your email address will not be published. Required fields are marked *

error: Content is protected !!