Localised ADAS Solutions Will Serve Indian Market Well: MathWorks

Localised ADAS Solutions Will Serve Indian Market Well: MathWorks

MathWorks Set Sights on Specific Localised ADAS Solutions for Indian Market
{data-social}

 


There is a significant amount of focus across the automotive industry on vehicle safety. Road traffic fatalities are indeed a grave concern for the entire automotive ecosystem, especially in the Indian context the concerns are huge given the country’s appalling road accident-related fatality record the country has registered over the years. MathWorks is a prominent simulation and testing solution provider of Advanced Driver Assistance Systems (ADAS) and autonomous driving systems. Auto Tech Review caught up with Shanti Medasani, Director (Autonomous Systems & Sensor Processing), MathWorks and R Vijayalayan, Manager, Auto Industry Field Application Engineering, MathWorks to understand more about the ADAS adoption journey in India.

India registers around 1.5 lakh road fatalities, which translates into a road accident-related death every four minutes across the country. This is where Advanced Driver Assistance Systems (ADAS) can play a pivotal role in mitigating road accident-triggered fatalities across the country. Shanti Medasani, Director (Autonomous Systems & Sensor Processing), Mathworks while acknowledging the importance of ADAS adoption in India, stressed the need for India-centric solutions. “Road fatalities are a big concern in India, especially on highways and implementing ADAS is replete with challenges. You can’t take solutions developed elsewhere in the world and just drop them in India and expect them to work. Given the country’s traffic density, it is imperative to have specific localised solutions. Infrastructure in India is improving in leaps and bounds, but there is a need for more intelligent solutions and this is where ADAS and connected technologies can ensure accident-free roads,” he said.

According to Medasani, no single solution can be a fit across the board. “You can’t have one single solution be it for highways, urban areas vs campuses. ADAS features such as Forward Collision Warning, Adaptive Cruise Control and Blind Spot Detection will be easy to deploy on highways while Parking Assist Systems can work well for campuses. For now, India can focus on these ADAS features,” he pointed out.

Touching on the ADAS implementation in rural areas, he said although the traffic density is lesser than urban pockets, ADAS still has to act, perceive and navigate as infrastructure is common for all environments be it urban, rural pockets or campuses.

R Vijayalayan, Manager, Auto Industry Field Application Engineering, MathWorks said the challenge lies with conducting physical testing exhaustively and this is where virtual testing and validation will hold the key. “There is a need for carrying out many generations of testing at extremely high speeds so that we know the system can tackle any situation and can safely navigate and complete its task. OEMs won’t have to reinvent the wheel, but leverage the expertise of other vendors and put together a solution they can stand behind,” he noted.

Vijayalayan shed light on the fast-evolving automotive ecosystem. “There is a great deal of collaboration happening between multiple stakeholders be it the auto industry, semi-conductor industry as well as the academia. Clearly, the OEM-supplier working relationship is evolving.”

The senior Automotive Application Engineering official noted that since a lot of sensors are deployed in ADAS, a lot of data will come into play. The challenge for OEMs and suppliers will be to gain insights from customer data, in terms of reusing the recorded data, he observed.

Vijayalayan stated that developing virtual testing grounds using simulation is something that would get OEMs excited as it can reduce physical testing and replicate dangerous scenarios. “OEMs are keen about developing virtual testing grounds. The functional safety norms are getting stringent and the focus is on how software can be developed quickly and how it meets the safety norms. MathWorks is developing one platform where you can develop and test your algorithms and ensure all modules are tested as it can reduce development time and increase confidence of the algorithms,” he remarked.

Medasani reckons ADAS-related sensors can be very expensive and all efforts must be undertaken to ensure frugality of sensors for the Indian market. “We need to see how we can ensure cost-effective sensors. For example, if we had cheap LiDAR sensors in India we would have ADAS already in our vehicles. We can also bring in a small subset of sensors for forward collision warning for it to work in India. Of course, vehicle-to-vehicle and vehicle-to-infrastructure can alleviate some of these concerns – you tell the environment where you are instead of someone else looking for where you are. So you need significant reliability of infrastructure. Further , robust high definition maps across the country’s road network can be leveraged - an area where India is seriously lacking,” he said.

Medasani said the industry-academia collaboration is on a strong footing in Europe, USA, China and Japan, while industry-academia collaboration is still in a nascent stage across India.

He stressed the deployment of digital twins in a virtual environment. “Everyone is creating digital twins off their vehicles and platforms and the time is right to deploy digital twins in virtual environments and see how one can well survive the tests.”

Medasani offered his perspective on the government’s move to make ADAS mandatory in all vehicles by 2020. “We should not look to bring out a solution for the Indian market, but rather focus on designing and building a solution that not only caters to the Indian market but is also used in global markets.”

Awareness about the benefits of the ADAS is equally important, feels Vijayalayan. “People in India are not even aware about the existing capabilities in vehicles – some don’t even know how to tilt their headlights.”

He dwelt on the essence of model-based designs. “We at MathWorks are focussing on facilitating model-based designs wherein people can converse with systems at a high level and use their digital twins to exhaustively test their systems in virtual environments. People can build their algorithms either for perception, navigation or planning and test it exhaustively and deploy it to any solution of their choice before moving to the real world, he explained.

Vijayalayan threw light on MathWorks’s open and flexible simulation platform, wherein customers can parameterise it according to their different needs – for example, the unique driving cycles in India. Every geography will have different requirements and this user-friendly platform will enable algorithm developers and testers to use the tool and try to achieve their objective.

Medasani wrapped up the conversation by offering his take on apprehensions in India that self-driving cars will take away jobs of driver in the Indian context. “The focus should not be on what we are replacing or who is losing jobs but offering a force multiplier that performs its task successfully and reliably and improves the quality of life, thus ensuring a secure road environment.”

 

 

 

 

 

Suhrid Barua
Author: Suhrid Barua
Suhrid has a predilection for anything concerning the auto industry and at Auto Tech Review he gives vent to this passion – he has a special liking for commercial vehicles and electric vehicle ecosystem. Suhrid tweets @ProsJunoon