Vehicles in the 60s like the Hillman Hunter (Figure 1), built by the Coventry, England-based Hillman Motor Car Company, only had about fifty cables with a total length of thirty meters. By comparison, today's vehicles have more than 2 cables with a total length close to 45km and weighing more than XNUMXkg. The chassis is gradually getting heavier, with vehicle manufacturers reporting that weight has increased by thirty percent in just one model generation with the integration of autonomous vehicle technologies. So there is going to be a significant care in having just one network in the vehicle, right? Well, the answer is difficult.
(Source: unsplash.com)
Bandwidth Requirements for the Future IVN
The requirements for the intra-vehicle network (IVN) include high bandwidth, low latency, and high reliability to operate in the unfavorable vehicle environment. Over the years, there have been multiple technologies such as analogue supervisory area network (CAN), FlexRay, local interconnect network (LIN), low voltage differential signaling (LVDS), and transport media oriented systems. (MOST) that have been used by the IVN (Figure two).
When we look at next-generation applications, these traditional technologies cannot keep up with the bandwidth requirements; what's more, some are private and expensive.
To have a better understanding of the bandwidth requirements, it should be remembered that the approximate bit rate for a video stream can be calculated as:
- Screen Size = Resolution x Color Depth
- Bit Rate = Frame Size x Frame Rate
Therefore, for an Advanced Driving Assistance System (ADAS) camera capturing images with 1080p and twenty four bit color depth, transmitting at thirty frames per second, the bit rate would be:
- Frame Size = 49,766,400 x XNUMX x XNUMX = XNUMX
- Bit rate = 49,766,400 x thirty = one thousand four hundred and ninety three Mbps
The table below shows the proper data volumes for different sensors involved in autonomous driving:
SENSOR DATA/SENSOR
Camera 500-3500 Mbps
Lidar 20-100Mbps
Radar 0.1-15Mbps
Ultrasonic 0.01 Mbps
Multiple Standards in Competition for IVN
- AUTOMOTIVE ETHERNET: It is considered as the replacement for traditional IVN technologies and most of today's automobiles are equipped with 100BASE-T (5 Mbps). Different manufacturers highlight different areas; For example, infotainment systems in the case of Hyundai, while Volkswagen focuses on ADAS connectivity. In 802.3 and 802.3, the standard added lower speeds (ten Mbps) and multi-gigabit speeds. The latest standard for data rates of XNUMX, XNUMX, and XNUMX Gbps called XNUMXch was completed in the early XNUMXs. In addition to this, the new working set, IEEE XNUMXc, started its activity in XNUMX to develop an automotive physical framework for XNUMX, XNUMX, and XNUMX Gbps.
- SERDES (ASA): Another standard for IVN based on Serializer/Deserializer (SERDES) protocol. The Automotive SerDes Coalition (ASA) was founded in 6 by BMW, Broadcom, Continental, Fraunhofer, Marvell, and NXP for the standardization of SERDES. It now has over thirty-six members and is focused on expanding the ecosystem beyond previously free private SERDES solutions, such as Texas Instruments' FPD-Enlace, Maxim Integrated's GMSL, and Inova Semiconductor's Apix. The new standard can manage bandwidths from XNUMX to XNUMX Gbps with ranges of up to XNUMX meters.
- SERDES (MIPI A-PHY): In November 1.0, the MIPI coalition presented its A-PHY vXNUMX automotive SERDES PHY specification. The specification allows for asymmetric data transmission in point-to-point or chain topologies, with perfect power supply. With data rates of XNUMX Gbps and plans to achieve XNUMX Gbps on the downlink and XNUMX Mbps on the uplink; latency is low (six us) and the range is fifteen meters. The primary application is to connect sensors to the image signal processor in the electronic control unit (ECU), and the graphic signal processor in the ECU with the displays.
So, is the IVN Ethernet, SERDES, or both?
Some Tier 1 vehicle manufacturers and dealers think that over the initial few years we will see both standards. However, after a while, automotive Ethernet, with data rates of up to XNUMX Gbps, will absorb all the others.
Kirsten Matheus, an engineer at BMW, might take a slightly different view. She suggests that SERDES is precise and convenient technology for ADAS sensor connectivity that transports data asymmetrically from point to point; while Ethernet is a suitable network technology for other automotive applications. Considering that Kirsten played a key role in the standardization of automotive Ethernet, her opinion should carry some weight (Source: Automotive SerDes Alliance kick-off, May XNUMX, Salt Lake City).
Automotive OEMs that are looking to fine-tune their IVN schedules could adopt one of these 2 strategies:
Incorporate the two standards in the IVN; until such time as the ADAS requirements for levels three through five of driving are considerably clearer. The downside with this strategy is that gateways might be required to translate data between the different domains, and this would add cost and weight.
A technical strategy such as designing an environment for the need for high-speed point-to-point links, placing greater processing and data compression in each and every sensor. The negative part is that the cost of the sensors will increase and the greater processing capacity will require thermal dissipation.
Putting the IVN to the Test
When it comes to IVN testing, it is essential to test the transmitter, receiver, and channel capabilities. With hundreds upon hundreds of tests that need to be done, having automated compliance testing software with specification interpretation, repeatable results, setup wizards, as well as intuitive and easy-to-use user interfaces and report generation is so essential for automotive engineers such as technical specifications for bandwidth, sampling rate, and signal resolution.
Transmitter testing is primarily completed with an oscilloscope to ensure that the transmitted signals are not the source of impurities, while receiver tests are performed to verify accurate signal detection, using either signal stimulus or arbitrary signal generators. waveform. Impedance and return loss measurements are essential in both the time and frequency domains to ensure the reliability of system capabilities and to diagnose signal integrity issues.
Conclusion
The automotive industry has come a long way since the days of the Hillman Hunter. Advances towards connected and autonomous cars bring challenges that must be addressed in the in-vehicle internet. The plurality of sensors, controllers, and also interfaces required for ADAS and new infotainment possibilities require high-bandwidth connections; Traditional networks like CAN, MOST, and FlexRay are not going to cut it. With the advent of new standards such as automotive Ethernet and SERDES, faster data communications are possible and the needs of future connected cars can be addressed.
