Kishore Ramisetty, Vice President & General Manager, Vertical Solutions and Services Group at Intel Corporation, has been associated with the organisation since 2013. This is his second innings with the technology giant as he has worked with the company as Director and Business Head, Innovative Products Group, from 2000 to 2012. Ramisetty has also worked with Trillium Digital Systems in the past. He was also the Chairperson of the CII Core Group on the Internet of Things from January 2015 to March 2017.
What’s the solution from Intel and the story behind it?
It started in October of 2017 at the Champions of Change event by Niti Aayog. We were talking about artificial intelligence (AI), Machine Learning (ML) and how they will disrupt industries. And our Prime Minister asked how AI impacts the life of a common man.
Our solutions are aimed to help the common people. It is based on a seminal paper released in Europe and Switzerland that talked about how six out of ten accidents that happen on roads can be avoided. We learned that if an alert is given to the driver, let's say, two seconds before an impending collision, he can avoid an accident because the driver can take evasive action. The research says 1.3 seconds is enough.
It took a bunch of engineers based in India who understand the road and traffic conditions in India and how drivers drive in India to develop the solution. But beyond that, it was three years of testing. We have seen almost 70% improvement in safe driving practices. In some cases, we have seen a 40% to 60% real accident reduction. In some cases, we have seen a 75% reduction in damage costs to the vehicle.
Interesting, but AI and ML require many data to perfect them. How did you solve that puzzle?
The three years I mentioned were the journey of data collection and perfecting the algorithms. The moment we fit a device in a truck or a car, it collects data. Today, we have three million kilometres of data from India and multiple more million kilometres in our testing phase. There were a bunch of test vehicles to which this device was deployed, and these vehicles were then sent on trips in the country.
Where do the calculations happen - edge or cloud?
All the real-time alert calculations happen locally on the devices (edge). Post-analysis takes place in the cloud. For example - 1,000 alerts on a trip from Bangalore to Pune are analysed in the cloud.
The cloud also analyses a driver's behaviour using data collected over a long period. Cloud and edge both work to ensure that the drivers drive safely.
Is it possible for a truck driver to tamper with these devices?
We have tamper detection mechanisms. When new technology comes in, it makes people feel cagey about it. We have tamper detection mechanisms in the device both in terms of drivers trying to obstruct the camera and inadvertently not maliciously trying to douse the devices.
Who's the majority adopter of the solutions you have showcased? Are fleet operators a focus?
Intel, along with Mobileye, has two businesses. The first business is to provide our product to OEMs. We reach out to OEMs via Tier-1 suppliers. In that line of business, we have almost 100 million vehicles on the road. This is generally for new vehicles.
The second division helps the aftermarket product deploy these solutions in older vehicles. Apart from the social causes, the fleet operators have two benefits: boost profitability because we reduce accidents, and hence downtime and profitability improve. It also creates safer drivers.
Then, where are most sales coming from - in CVs or the car segment?
The volumes are in the passenger car. However, our focus in India is on the commercial vehicle industry. The reason is that almost 60% to 65% of our accidents happen on highways in India. So the need for this technology is first in those commercial vehicles. That's why we have focused specifically on commercial vehicles.
Can these solutions be simply plugged and played in existing trucks? What about very-old trucks?
The solutions do not work with legacy trucks as they do not have CPUs. These solutions work with trucks which are BS III and above. Older trucks are not generally used for longer-hauls. They are mostly deployed on shorter routes.
Do you see the existing telematics OEMs used as a competitor to you?
Our solutions are complementary to the telematics used by OEMs. We create a unified dashboard of both where we have input from vehicle telematics and then add the aspect of driver monitoring, which is about vehicle safety. The solution tells much more than fuel usage as it informs about driver safety and behaviour.
What's in it for OEMs to go with your solutions?
Every OEM is preparing for the future. They need to understand everything about the fleet, how the driver is driving, how the vehicle is being driven and every aspect of it. OEMs have only one component of it, which is that they understand the fault codes and engine. We have the rest of the pieces of it.
A lot of fleet owners do not want to see two different sets of data. Intel partnering with OEMs will help them give unified data to the fleet owners.
Can the data you collect help an OEM?
Yes, it's possible; depending on the business model, Intel can front-end the customer by taking the OEMs data into its services, or we can provide data to an OEM. The OEM then can provide the data to his end consumer. Both conversations are happening right now. The goal at the end of the day is for the customer to be happy.
What is the difference in cost for OEMs and the aftermarket?
OEMs mean business in huge numbers. After signing a contract, every truck that rolls out of the line would have the solution implemented. It would be very hard to quantify it between the two as the numbers are entirely different.
Related Read