MOST Forum hears research report: Gbit capacity EM propagation in plastic tubing
Since last year’s MOST Forum several carmakers have announced new models with MOST technology incorporated. Among them are Volvo’s latest top-end SUV XC90, BMWs 2 Series Active Tourer and the Audi TT. In the latter case, a MOST INIC (Intelligent Network Interface Controller) is used in a somewhat unusual and innovative way: It is part of Audi’s Virtual Cockpit System and thus in a domain that is not associated exclusively with the infotainment – probably an industry first. Meanwhile, in some 180 vehicle models worldwide, MOST is the underlying networking technology for the infotainment. While MOST networks are found mostly in upmarket and luxury cars and SUVs from carmakers such as Audi, Bentley, BMW, Daimler, and Rolls Royce, the technology has also arrived in the segment of more affordable vehicles. An example is the Smart Forfour as the first subcompact vehicle with MOST-based infotainment.
At the MOST Forum, it appeared that the MOST technology has found its niche where it will be difficult to attack by its fiercest competitor, Ethernet AVB. “Certainly, Ethernet has a place in the world: Where smaller amounts of data with uncritical real-time behaviour need to be processed, packet-oriented technologies like Ethernet have their advantages”, said Henry Muyshondt, Administrator of the MOSTCO, the group of companies that have gathered to develop and support the MOST technology. “However, when it comes to transport [of] large amounts of streaming data for audio, video and camera signals, MOST will continue to be the superior solution.”
At the event, techniques to test and evaluate MOST-based systems took centre stage. Software options to integrate MOST systems into Autosar environments were also detailed by experts from Bosch subsidiary ETAS as well as from MOSTCO.
But in which direction will the MOST technology develop in the future? Where are the priorities – will MOSTCO put the focus on additional functionality to turn it into a networking platform that can be deployed in application fields beyond infotainment? Or will the ever-increasing bandwidth requirements of future applications (camera signal processing for automated driving functions, for example) set the pace for the developers? In the short to medium term, users should expect more features and functions. Muyshondt said. “In the current market situation, this is more important than additional bandwidth”. The MOST group spokesman added that in its current version, MOST technology also supports multiple concurrent data streams and offers a full-fledged Ethernet channel. It also has a coax physical layer that does away with the cost disadvantages of optical data transmission. Nevertheless, in the longer-term future, the bandwidth demand will return and enforce new developments, Muyshondt acknowledged. “Though MOST50 (with a bandwidth of just 50 Mbps) is still our bestseller, new applications such as uncompressed camera signals and ultra-high definition video might require data paths in the Gigabit range,” he said.
Under this perspective it was no coincidence that some presentations dealt with ways to push the bandwidth limits of future MOST generations. Conrad Zerna from Fraunhofer Institute for Integrated Circuits (Fraunhofer IIS) showcased a prototype version of a 10 Gbps Physical Layer for Single Twisted Pair cabling (and thus the same cost-effective type of wiring as promoted by the Ethernet AVnu Alliance). The goal of the developers is transmitting this this amount of data across a distance of 10m at a latency of less than 10 microseconds and a power consumption of less than 1W in the TX/RX combination. Applications could be in-car data backbones or automotive MIMO arrays. At present, the team has devised a bandwidth of 7.9 Gbps across a distance of 14m; Zerna hopes that through further optimisations it will reach the goal of 10 Gbps by end of May.
Another presentation went significantly beyond the 10 Gbps solution from Fraunhofer, albeit it was at present purely a research project, far from commercialisation. Researchers from the KU Leuven (Belgium) showed an alternative technology to copper and optical fibre which enables wide data bandwidths well inn excess of 10 Gbps. The team led by professor Patrick Reynaert and Wouter Volkaerts succeeded in transmitting RF through plastic fibres – just the electromagnetic waves, no transformation to optical pulses is required. Instead of LEDs and photodiodes or laser emitters (VCSEL) and PIN diodes, the team uses cost-effective CMOS circuitry and tiny copper antennas to transmit and receive the data. The antennas could easily be implemented at chip or package level, Volkaerts said. “Our research shows that a data rate times distance product of 100 Gbps.m is possible”, Volkaerts claimed. Which means, for instance, 10 Gbps over a distance of 10m; at shorter distances the data rate could be even higher.
The beauty of this approach is its cost-efficiency. Instead of optical fibre (or even Plastic Optical Fibre) it uses cheap plastic tubing or similar material. “We bought ours in the hobbyist store”, Volkaerts smiled. Despite the simple material, the technology exhibits high robustness with respect to electromagnetic interference and does not require extremely precise, expensive mechanics to couple the link with the circuitry.
While the Leuven Scientists pointed out that at present this is just a research project, the interest from the Forum attendees was high. Also MOSTCO Administrator Muyshondt said that perhaps this technology could some day the base for a future MOST physical layer.
What else? Talks with attendees showed that perhaps the Ethernet camp’s much acclaimed cost-effective unshielded twisted pair might fall short of being the best solution for high-bandwidth data networks in the car. "Even a slight strain or kink of such a cable leads to heavy signal distortion", complained an electronics designer who asked not to be identified. "MOST might be somewhat more expensive, but at least it works reliably".