A fleet of machines, electrically-powered and controlled by artificial intelligence (AI), guided by an array of sensors, radar and LiDAR: vehicles of the future may bear little resemblance to passenger cars as we know them. And the transition is well under way, prompting an equally radical shift in design and engineering.
Tools to help
MathWorks has been helping automotive companies with their development efforts by providing solutions on technical computing and model-based designs. Earlier this year, the Natick, Massachusetts-headquartered mathematical computing software company released a new product called Automated Driving System Toolbox, which provides algorithms and tools for designing and testing advanced driver assistance systems (ADAS) and autonomous driving systems.
Engineers face a host of challenges in developing such systems, including long development cycles, system and algorithm creation, and testing and verification. Toolbox helps with all of that. “You can automate ground-truth labelling, generate synthetic sensor data for driving scenarios, perform multi-sensor fusion, and design and simulate vision systems,” explained Kishore Rao, Country Head and Managing Director for MathWorks in India.
Data scientists combine three types of knowledge. First, they have domain expertise. Second, they understand computing. Third, they understand how to apply statistical and mathematical analysis to their problems. It is very rare to find people who have all three
The company also has tools to help with a host of emerging megatrends, including the move towards tighter emission standards and full vehicle electrification. Companies face increasingly complex controls and diagnostics due to the proliferation of sensors and actuators in vehicle systems. At the same time, more stringent emission standards demand precision in control and calibration. In 2016, MathWorks released a product called Powertrain Blockset, which provides fully assembled reference application models of automotive powertrains, including gasoline, diesel, hybrid, and electric systems. Customers can use it for design tradeoff analysis and component sizing, control parameter optimisation, and hardware-in-the-loop testing.
The company is also preparing for the many opportunities made possible with Big Data and predictive maintenance. In contrast to conventional preventive maintenance, the schedule for predictive maintenance is not determined by a prescribed timeline. Instead, it is determined using analytic algorithms harnessing data collected from equipment sensors.
When it comes to tackling the automotive design challenges of today and tomorrow, a host of advanced skill sets will be required. “Data scientists combine three types of knowledge. First, they have domain expertise. They are experts in the field in which they work,” said Rao. “They know the engineering or the science behind their projects. Second, they understand computing. They know the basics of coding, data management, and computing infrastructure. Third, they understand how to apply statistical and mathematical analysis to their problems. It is very rare to find people who have all three types of knowledge, so that makes them difficult to hire. Almost every week there are articles written about the shortage of data scientists, and it is recognised as a global issue.”
The skills shortage
Players across the automotive spectrum are struggling to acquire individuals with new skills in emerging areas such as data science but also software, deep learning and AI. “This is a concern for the industry – how do you secure the talent in these new areas? How do you develop the skills or bring in the skills?” asked Rao. Because MathWorks provides the tools and the knowledge on how technology can be used in these new areas, it believes it can play a special role in this challenge.
The IT services tend to dominate the hiring season in any engineering institute. Many skilled individuals go to services companies and take on a programming profession
One way for companies to re-skill their teams is to send their engineers to courses on data analytics, deep learning, etc. They can also hire a team of experts, but the shortage of qualified individuals is a real obstacle. “We are a big proponent of equipping the domain expert and the engineer with tools that will allow him to learn these skills very quickly. That is something that has been a big focus in our tools in the last two years, and something that customers really value,” observed Rao. Enter MATLAB – it essentially provides an interactive environment and a set of tools to enable domain experts to become data scientists, developing custom predictive models from engineering and business data. Through flexible deployment options, production-ready models can be integrated much more quickly into business systems and embedded devices.
“Rather than hiring costly data scientists who may not have the correct domain expertise, you can look to MATLAB users to apply statistical methods and machine learning to solve engineering problems. These engineers and scientists already have the domain expertise, so they are able to quickly determine whether an analytic technique is going to be useful,” Rao explained. “Rather than just hiring the data analytic expert from outside, who might not have the domain knowledge, there is a lot of value in re-skilling the engineers who have the domain knowledge.”
Outside of industry
That works well for people who are already working with automotive companies, but attracting players to the industry in the first place is another challenge. MathWorks takes a two-pronged approach. “Not only do we want to enable existing engineers in industry to skill up faster in these areas but we also want to work with academia to help students to learn this faster. We need to look at people in academia because they will be the engineers of tomorrow,” noted Rao.
India is known for its IT services. Our intent, when we work with engineering colleges, is to direct them to the core industries, including automotive
MathWorks puts considerable focus on academia and is active on various student competitions, such as Mahindra Rise’s autonomous vehicle challenge and Robocon India. This latter competition encourages engineers to tackle problems in the fields of robotics and AI.
MathWorks has been a sponsor for Robocon India for the last few years and offers participating teams free access to MathWorks tools along with a chance to win the MathWorks Robocon prize in addition to the Robocon award.
“We work with the students participating in these competitions, provide them tooling, software and training. It gives them the chance to start using these tools for developing products,” he commented. “We try to catch these kids early and make sure that they are learning industry standard tools, new domains and how to ramp up.”
We are a big proponent of equipping the domain expert and the engineer with tools that will allow him to learn these skills very quickly. That is something that has been a big focus in our tools in the last two years, and something that customers really value
These activities are more than simple education outreach. For MathWorks they represent a strategic focus. In fact, Rao describes the focus it places on academia as ‘disproportionate’. Globally, just 10% of the company’s revenue comes from academia, while 90% comes from industry. “And yet we focus much, much more on academia – about 30-40% of our mindshare and resources go there. That’s because these people are going to be the future engineers.”
So far, the strategy is paying off. “People who are coming into the automotive industry from academic institutions are coming in with some level of MATLAB knowledge and they are able to ramp up faster once they join the industry,” he noted. There’s also a greater diversity of talent than in the past. “In India, it is quite diverse, much more so than when I studied engineering in the 1980s,” he pointed out.
Despite the progress, attracting individuals with the right skills in certain areas remains a significant challenge, not only for MathWorks but for all automotive companies. Rao describes the hunt for talent in general as “very competitive”. A big part of that challenge in India is due to the lure of the IT services segment. “The IT services tend to dominate the hiring season in any engineering institute. Many skilled individuals go to services companies and take on a programming profession,” said Rao. “India is known for its IT services. Our intent, when we work with engineering colleges, is to direct them to the core industries, including automotive.”
This article appeared in the Q4 2017 issue of Automotive Megatrends Magazine.