MathWorks | Data Analytics Within The Context Of Model-Based Design

MathWorks | Data Analytics Within The Context Of Model-Based Design

{data-social}

Model-based design has been an important part of product development across industries, including the automotive and aerospace industries, and MathWorks has been the market leader offering model-based design solutions. At the MATLAB Expo 2016, we met Richard M Rovner, Vice President, Marketing, MathWorks (L) and Kishore Rao, Managing Director, MathWorks India Pvt Ltd (R) to get a lowdown on the new solutions the company is bringing to the Indian market.

ENGINEERING-DRIVEN ANALYTICS

Microprocessors in vehicles generate enormous amount of data, and MathWorks sees such engineering data as a key driver to enhance analytics, which have become pervasive in all industries. MathWorks continues to invest in the MATLAB and Simulink tools to enhance engineering-driven analytics workflows that would result in new capabilities of Big Data processing, new algorithms for machine learning, and computing.

Fleet data analysis, calibration and automated driving algorithms are some of the most sophisticated uses that analytics are put to. Deep learning and predictive-models for driving algorithms are expected to have a deep impact on advancements in automated driving, as well as in driver assistance and active safety, noted Rovner. He added that analytics also includes computer vision and co-generation, which can be addressed within the context of model-based design.

The most significant product addition to the company's flagship platform was Simulink Test that can be used to systematically verify and simulate models, said Rao. The Simulink Test not only helps model simulation in desktop simulation, but also adds to the hardware-in-loop (HIL) simulation capability.

Rao said MathWorks is especially focussing on developing four technologies that are most relevant to the automotive sector. The Automatic Production Code Generation technology is used by OEMs to generate codes and use the same in production programmes. MathWorks is also increasing the use of System-Level Simulation, where physical modelling is used mainly for the behaviour and designing of physical systems. ADAS is being developed, especially since automakers are incorporating this technology in developing automated and autonomous driving vehicles. The company is also investing in computer vision technology that is critical for ADAS.

Meanwhile, Model-Based Calibration technology generates optimal calibration tables that are used by automotive companies to meet fuel economy requirements and emission regulations. This technology provides benefits in the form of reduced cost and man hours, when compared to calibration using a dynamometer and can be used to simulate the test bench itself. Another main advantage of model-based calibration is that a problem can be specifically defined to see its outcome and effect on the engine. The move towards early adoption of Bharat Stage VI norms in India has renewed interest within domestic OEMs for such technologies, noted Rao.

TRENDS & CURRENT REQUIREMENTS

With automotive manufacturers taking a comprehensive approach to model-based design from beginning to end, verification and validation (VnV) have become important parts of the workflow. In addition, one of the most visible trends in the automotive industry at present is in the development of ADAS for automated driving, where there are a number of challenges, noted Rovner. Systems including computer vision, learning systems, ground truthing, ground truth labelling and sensor fusion pose a range of challenges that the company works to resolve. Other trends include model-based design to address fleet data analytics and compliance to user and safety standards. There are a few standards that mandate the use of model-based design, which will drive the quicker adoption of such technologies, he noted.

TECHNOLOGY ADOPTION

The use of MATLAB and Simulink in model-based design of domestic OEMs and suppliers has reached a tipping point over the past couple of years, Rao said. Companies recognise the benefits of using model-based design and are building competency for use of these technologies. The level of awareness and need for model-based design has grown and been appreciated with use in real projects. While OEMs still work with suppliers on their design front, it has changed significantly, with more of a co-development approach being employed, which will reduce the time to market.

On the other hand, design work carried out by MNCs in the country isn't very different from their global developments. Their development work is driven by global projects, and local teams collaborate with their larger worldwide engineering teams, said Rao. He claimed MathWorks solutions help OEMs and suppliers to reduce re-work, cut down coding errors, and aid efficient development of variants.

Rovner said MathWorks will continue to make strategic investments for toolboxes in its core platforms of MATLAB and Simulink, in order to continuously make improvements to the entire workflow. It is also investing in resources to help customers adopt its technologies, in terms of online training and examples, as well as investments in people for application engineering, training, consulting and pilot engineering.

Text: Naveen Arul