The Electronic Design Automation (EDA) industry is witnessing a growing set of challenges, with technologies like driver assistance systems, autonomous driving and connected vehicles gradually coming into the picture. Cadence Design Systems has been in this field for a number of years and continues to offer solutions pertaining to chip design for electronic systems.
In an interaction, Apurva Kalia, Vice President - R&D, Advanced Verification Solutions, Cadence Design Systems, spoke about the current state of the EDA industry, new technologies for vehicle safety and security, and autonomous vehicle challenges. Kalia also touched on the role that automotive electronics will play in the future, and discussed the opportunity that Cadence has in offering solutions to the automotive industry.
A fact that has been holding up in the EDA space, as well as the electronics space for a long time is Moore’s law, which states that the density and complexity of chips doubles every 18 months. This trend of increasing chip density has driven innovation that is seen in electronics and EDA industries. This doubling of density has resulted in the reduction of electronic geometries from 120 nm to just 7 nm, for which different types of tooling and innovation are required.
For the automotive industry, vehicle electronics have been important in the context of ECUs and infotainment, both of which are considered as ‘simple’ electronics. The ECU is a mission-critical component, but the electronics are relatively simple, whereas the electronics in an infotainment system are complex but not mission-critical. Now, complexity and mission-critical electronics are getting integrated when it comes to driver assistance systems, leading up to ADAS.
ADAS involves extremely complex electronics, and the chips for these systems are more complex than any other chip that has been designed so far, noted Kalia. He said that these complex chips can be manufactured in a cost-effective way only if the density is high. Kalia said that Cadence’s aim is to give customers tools for EDA solutions, providing solutions for development, designing and verification. The challenge is to develop methodologies that will allow designing and verification of a chip so complex, within that short period.
Another major challenge for the EDA industry is in making chips that not only are complex, but also secure. All this calls for new tools and verification methodologies that will not only allow users to verify that their chips work, but also prove that they’re safe. Cadence allows users to design chips and get them certified to be compliant with required safety standards, noted Kalia.
Machine Learning and Data Analytics are an integral part of the ADAS equation, noted Kalia. He said that the ADAS challenge is essentially in hardware, where the chip needs to be designed to carry out a high level of processing. One part of the problem is that the chip needs to be able to analyse the environment around the vehicle in real time and provide input with a very miniscule latency time. This will require a huge amount of data to be processed in a very short period of time. The reasons for ADAS becoming possible are, first, the possibility of high density and complexity of chips, and second, the development of machine learning.
In terms of the challenge concerning autonomous vehicles, Kalia said it comes in two parts – technology and non-technology challenges. The technology challenge is in electronics and software, since the software needs to also react to changes in the environment within an extremely short period of time. Additionally, processing capabilities need to be sufficient to process all the data received from 18-20 on-board sensors, which provide up to about 4,000 Gigabits of data per hour.
Then there are non-technology challenges, which include regulations and liability. In terms of regulations, there is a gap in identifying the right party to certify an autonomous vehicle, along with providing the conditions necessary to carry out this certification. Little thought has been given to this process of defining the requirements of autonomous vehicles, observed Kalia. Another part of the non-technology challenge is in identifying who the liability lies with, in case of an accident happening with the autonomous vehicle.
Traditionally, electronics like ECUs and infotainment systems have been a small part of the value of the car, but with ADAS, electronics will become a large part of the total vehicle value, noted Kalia. Some OEMs expect full ADAS with Level 5 autonomy to comprise 45-50 % of the total value of the car. Kalia said that an interesting trend is that OEMs want to be able to control the ADAS sub-system, due to the high value. This is leading to OEMs taking up the role of Tier-1s, in the sense that they want to be able to develop technologies like ADAS in-house. This is also pushing Tier-1s to take up the role of their suppliers, and begin providing the electronics for ADAS to their OEM partners.
This trend of changing roles between OEMs and Tier-1s brings its own set of challenges, while also providing further business opportunities for Cadence. While the challenge is that a completely new set of customers need to understand the domain, the opportunity can be utilised to help them understand the domain and offer appropriate solutions. It is a similar case when it comes to the development of electronic technologies for electric vehicles. Kalia noted that Cadence is offering solutions from its mobile phone portfolio relating to low-power designs for electronics, in order to help in developing solutions for electric vehicles, which also require reduced energy consumption.
The next decade for the automotive industry would almost be like the 1990s were for the communication industry, in terms of the development that will be seen. Additionally, the value in automobiles is shifting towards electronics and software, which will continue to provide Cadence with opportunities in creating solutions for future needs pertaining to electronic components.
TEXT: Naveen Arul