Home > Analysis > Model-based design crucial for securing ‘intricate system interplay’

Model-based design crucial for securing ‘intricate system interplay’

The amount of data produced by vehicles is set to grow dramatically with the rise of the connected car. This data can be leveraged effectively with the help of model-based design. By Michael Nash

The design processes for vehicles continue evolve. Laboratories are no longer full of full-size vehicle prototypes and rough sketches splattered across reels of paper. Instead, computer software has become capable of providing highly detailed and accurate simulation. This model-based, simulated approach to vehicle design could be vital as the industry faces increasingly tough requirements and demands, particularly those centred upon safety and connectivity.

From powertrain to safety
Mathworks, a mathematical computing software developer, hopes to accelerate the pace of engineering and science by providing technical computing software that can be used across a variety of industries.

Speaking to Megatrends, Kishore Rao, Managing Director, Mathworks India, explained the parameters governing demand for model-based design in the automotive industry.

“At Mathworks, we look at automotive trends, like emissions, fuel economy and safety,” he explained. “We then aim to provide a platform for our customers to address these three drivers. Our solutions, like Matlab and Simulink, can become an integrated platform for data analytics and model-based design, allowing companies to address the challenges that they face.”

Matlab is used by over a million people around the world for machine learning, computational finance, control design, signal and image processing. It is also a common tool for the medical, energy, cellular, spacecraft and automotive industries.

While Matlab uses a coded language, Simulink provides a block diagram environment for model-based design. It is integrated with Matlab, and allows users to incorporate Matlab algorithms into models and export simulations to Matlab for additional analysis.

Traditionally, these model-based design tools have had “tremendous success” in the field of powertrain development, Rao explained: “I think the reason powertrain was the initial area to adopt model-based design tools is primarily due to increasingly stringent CO2 regulations. Companies realised that they could no longer reach the targets with traditional systems, which included fine-tuning the powertrain on the bench. They turned to advanced control concepts in the hope of getting more out of the fuel they inject. This paved the way for the use of models, and simulating the entire system with software tools.”

Looking ahead, Rao expects this success to continue as the electrification of the powertrain increases system complexity. “There’s an even stronger need to simulate and carry out sizing exercises.” He also referred to the development of “model predictive controls” – these, he explained, are sophisticated engine control systems that “continue to push the envelope on powertrain design.”

While he thinks that the powertrain field will remain the biggest user of model-based design, Rao revealed that the largest growth area could be vehicle safety thanks to the rising popularity of advanced driver assistance systems (ADAS). “There used to be an exclusive focus on active safety, but now with the rise of driver assistance, there is a big push towards passive safety. Model-based design tools can be used to create and validate these complex systems, and validation is a major component of these systems.”

Like many other experts, Rao thinks there is a strong link between ADAS, the realisation of autonomous driving and vehicle safety. While most OEMs have already started to introduce ADAS in their portfolio, only a handful of companies like Tesla have shown their autonomous vehicle capabilities.

“There are varying levels of automation making inroads in the industry,” Rao affirmed. “All of these are very tightly linked to safety. It is no longer one domain or one subsystem that is designed to tick the boxes. For example, as you go into ADAS, there is an intricate interplay between the vision sensing and the braking system. This communication needs to be actuated to be able to predict the scenario unravelling ahead, and to provide the necessary means to mitigate or avoid a collision.”

The reason powertrain was the initial area to adopt model-based design tools is primarily due to increasingly stringent CO2 regulations. Companies realised that they could no longer reach the targets with traditional systems, which included fine-tuning the powertrain on the bench

With data and connectivity comes complexity
Increased ADAS adoption is not the only trend providing opportunity for model-based design, Rao continued. The rise of the connected car has taken the automotive industry by storm, and many experts predict that vehicles will only become more sophisticated, smart and connected to keep up with consumer demands.

A recent report from SNS Research predicts connected car services to bring in US$14bn in annual revenue, marking a CAGR of 31% between 2016 and 2020. As Rao points out, “New areas of vehicle connectivity generate huge amounts of data, and there’s an opportunity to use this data by sharing it between different systems to improve their performance. Model-based design allows companies to leverage data and make better decisions on product design and refinement.”

Data analytics, he added, is an already huge area that will play an increasingly important role in vehicle design for every OEM. This will be evident in the design of all systems and vehicle components, be it in the area of safety or powertrain.

“Just to summarise, powertrain continues to be the biggest user of model-based design, but we see ADAS and safety as the big emerging area. Data analytics impacts all of these areas, and so leveraging it in an efficient and performance-optimising way will be the key trend,” Rao noted.

With both passive safety and vehicle connectivity playing an increasingly important role in vehicle design, Rao thinks the amount of data could lead to a number of issues in vehicle design. For example, he referred to a scenario whereby each and every sensor on the vehicle must be able to communicate and transfer data to the other, otherwise the safety of the driver and occupants is compromised.

“Everything is very complex today,” he emphasised. “with widespread use of electronics and software, and an increasing number of sensors in the vehicles. There is also considerable interplay between different systems, and a huge amount of data to juggle. The model-based design environment provides automotive companies and their engineers with the ability to conceptualise systems at a high level, simulate their interplay and ensure that the performance they are getting addresses their targets.”

Additionally, the need to secure this data and connectivity against cyber security is a growing concern for vehicle designers. Rao suggested that Mathworks is focusing more on this area than ever before, particularly in the wake of a small but growing number of incidents involving researchers successfully hacking into connected cars and taking over primary control.

“We’re looking at cyber threats with a view to securing different systems or algorithms that have been designed for the vehicle, and how they can perform in a certain way against hacking or against misuse,” Rao explained. “It’s a relatively new area for us, and one of the key starting points is the securing of the embedded systems and code. If the model-based design approach is used to simulate and test all conditions before using automatic code generation, then we can conclude with a high degree of confidence that the code does not have bugs.”

Securing data and code, Rao concluded, will be a crucial enabler not just for the rise of the connected car, but also for the increased use of ADAS and the realisation of autonomous driving.

This article appeared in the Q4 2016 issue of Automotive Megatrends Magazine.