What happens when technology from the fast paced communication industry makes a move into the traditional automotive industry? Semiconductor marketers and even the automotive industry are talking about revolutionary changes inside and outside the vehicle.
What kinds of changes? Ethernet and sensors. There’s a lot of excitement and enthusiasm over the prospect of cars with Ethernet networking capabilities and multiple ports for streaming video, driver-assist cameras, real-time diagnostics and autonomous driving. Ethernet is touted as being a faster, and ultimately cheaper, network solution for the operational information collected by sensors as well as providing more bandwidth for the infotainment needs of each passenger.
After attending DAC last week, I was quite bullish myself. Companies such as Synopsys are reporting significant design activity focused on automotive applications. Synopsys offers their DesignWare ARC SEP Processor for ISO 26262 safety compliant solutions as well as a sensor IP subsystem for small, low power devices.
Ethernet is already ubiquitous in the communication world, providing not only experienced design knowledge but also large-volume manufacturing. Large volumes lead to economies of scale and lower costs compared to other automotive network options. Cadence has a long history of designing Ethernet IP and developing standards which will make adoption in the automotive industry much smoother.
Broadcom, Freescale, Marvell, Micrel, and Renesas are just a few of the companies who already have design wins or offer solutions for Ethernet applications in the car. As expected, luxury vehicles such as the BMW X5 already offer 100Mbps Ethernet to connect driver-assistance cameras. But the prospect for Ethernet to trickle down to mid-priced cars is also happening. In 2012, Hyundai and Broadcom announced a joint development agreement to integrate infotainment, telematics and ADAS (Advanced Driver Assistance Systems) using Broadcom’s BroadR-Reach® Ethernet technology. But we haven’t heard anything since then.
After all the excitement at DAC, I returned to Phoenix to attend ON Semiconductor’s analyst meeting. Automotive sales comprise 27% of ON’s total revenue. Executives at On Semiconductor aren’t quite as enthusiastic about the rapid or broad adoption of Ethernet as I was.
This is truly a difference in perspectives. What is revolutionary to a semiconductor company dealing with cell phone technology that changes every year or two is completely different to the automotive industry that deals with design cycles of over five years. Design cycles for automotive used to be even longer. Will the automotive industry adopt new technology with the same dexterity as the smartphone market? Probably not. There are safety standards and legal issues to contend with, as well as a host of other challenges. Will Ethernet totally replace CAN or LIN as the network solution? Probably not. CAN and LIN will continue to fill a role for years to come.
Slow and steady is the mantra for new technology adoption in the automotive industry. But change is occurring. The adoption of automotive Ethernet solutions will be slower than expected for most semiconductor manufacturers but much faster than what most traditional automotive suppliers are expecting. I’m still enthusiastic about new technology in cars and will even bet that my next vehicle will have Ethernet and more sensors.