Publication Name: https://economictimes.indiatimes.com
Date: October 2018
For most parts of the brief trial around the campus, HCL Technology’s autonomous car —a retrofitted Ford sedan—felt like it was being steered by an upstart driver cautious around the sharp turns. The autonomous touch emerged when a designated human obstacle walked slowly on to the path and the brakes plunged abruptly several feet ahead of him. After his passage, the car drove itself on.
HCL is testing the vehicle retrofitted with autonomous capabilities such as sensors, cameras that will help the car see all around, new-technology braking systems, and a self-learning navigation platform that will collate scenarios through simulations.
The car will never hit the market, but like its rivals in India’s IT services industry such as Cognizant, Tata Consultancy Services, and Tech MahindraNSE -0.84 %, HCL is building these capabilities to showcase to potential customers.
A CB Insights report says 46 companies including carmakers such as General Motors, Audi and Ford, besides components makers such as Bosch, Continental and Delphi, are working on autonomous cars or enabling technologies for such vehicles.
“The challenge is to hit the right balance between costs and features, and we are a long way from there. Initially, we are trying to come up with specific use case of the algorithms, especially unique ones, which are part of the maturity curve,” GH Rao, president of engineering services at HCL, told ET.
In a market driven by competitive go-to-market standards, HCL is rapidly feeding the algorithm with videos taken on roads, citywise, and in-lab simulations to make the software robust. Some of its customers are already being engaged with some autonomous capabilities like auto-braking systems and other Intellectual Property-armed services.
“Another challenge is bringing an awareness to the car. As of now, we are developing capabilities in communications, remote diagnostics, over-the-air upgrades...
We are also doing a lot of work on (safety systems)—lane warning systems, linking mirrors to project a 360-degree view, and so on.”
TCS has a team of over 800 engineers working with both established carmakers as well as startups in Silicon Valley that are rushing to build autonomous vehicles.
“Every car company is working on autonomous cars. While they are building products for the short-term, as part of the long-term initiatives, they are spending significant amounts of money on these platforms,” said Regu Ayyaswamy, global head, engineering services and Internet of Things, at TCS. “It is translating into more business for us.”
Even as technology companies productise their technologies, research is underway to maximise autonomy in driverless cars. IITMadras is working with American technology company Intel and German firm Continental, which trialled a driverless car in Frankfurt last year.
Professor Ravindran B, who heads the Robert Bosch Centre for Data Science and Artificial Intelligence at IIT-Madras, said: “We have worked on building AIbased decision-making algorithms that allow us to adaptively learn controllers for driving. In particular, our group has explored the question of learning from demonstrations. These are largely laboratory experiments and have not yet been translated into real systems.”
With Intel, researchers from IITMadras and IIT-Kharagpur have contributed to building a simulator for training autonomous vehicles. The research centred around building a simulated environment much like a real-life traffic situation—where multiple drivers react autonomously, as opposed to existing algorithms were a single agent works on its own while others are just programmed to react in a certain way.
Ravindran envisages cars with some autonomous capabilities before full-fledged driverless cars hit the roads: “In India, we are more likely to see Autonomous Driver Assist Systems (ADAS) in the near future, and fully autonomous vehicles might be some way off—partly due to the unstructured driving challenge, and partly due to the government’s opposition at a policy level. We could see autonomous drones/cars in controlled environments before we get to cars.”
In the future, software companies see a shift from people wanting to own cars—expected to impact vehicle sales—to “purpose-built vehicles” operating in a shared economy, say experts.
Cognizant, another software major invested in the field of autonomous vehicles, tied up three years ago with the Pune-based Automotive Research Association of India. ARAI is a collaborative effort between automotive corporates and the heavy industries ministry. As its clients, Cognizant has original equipment manufacturers that are making platforms based on AI and machine learning technologies for autonomous cars. Among its many services, Cognizant helps carmakers in a systems engineering concept called Verify and Validate: to virtually help cars identify objects in their range of vision accurately. Queried about the market opportunity for Artificial Intelligence products, Madhuraj Jadav, Vice-President of connected products at Cognizant, said: “AI-based autonomy is being developed independently by many players in the industry across the globe.
However, the algorithms and solutions being developed are proprietary and not a standard.
Therefore, we perceive that the overall opportunity is significantly higher with high growth potential. We have built expertise within Cognizant and have developed accelerators to explore these opportunities.”
V Vinay, founder of ATI Motors, which is building an autonomous cargo vehicle in India, said that while the opportunity to build technologies is there for companies in the country, the biggest challenge is trained manpower.
“We don’t have enough people who understand engineering concepts for autonomous vehicles at scale that are required for services companies,” said Vinay.
Tech Mahindra, another player in the autonomous car space, foresees a market opportunity of $5-6 billion for domestic engineering services companies, growing at a clip of 35-40% from the current market size of $700-800 million.
To capture the opportunity, Tech Mahindra has over 1,000 engineers in its Autonomous School of Excellence and plans to add nearly 2,000 more in the next two months.
“In my interactions with customers in Germany and China, I see that the immediate need is at least 20,000 engineers with knowledge in automotive software space,” said Karthikeyan Natarajan, global head for integrated engineering services at Tech Mahindra. “Nearly 70-80% of the work going on in autonomous vehicles is in third-degree autonomous vehicles—partially autonomous wherein the driver can take his feet and hands off and the car keeps going.”
Tata Elxsi has built an autonomous vehicle middleware platform, Autonomai, designed to help OEMs and system suppliers build, test and deploy autonomous vehicle applications. Elxsi is also jointly working with Hyundai to develop a simulation environment to test scenarios for autonomous driving.