Twitter icon
Facebook icon
LinkedIn icon
Google icon
Reddit icon
StumbleUpon icon
Del.icio.us icon

As complexity intensifies, in-vehicle networking gets a radical rethink

Added to IoTplaybook or last updated on: 08/20/2021
As complexity intensifies, in-vehicle networking gets a radical rethink

To keep up with the demand for more processing power and blazing-fast data transmission speeds, the auto industry is moving toward a zonal electronic/electrical (E/E) architecture — a near-complete departure from the way vehicles are designed today.

The Internet of Things may be a complex data communications environment, but it’s got nothing on today’s vehicles that cram an entire industrial IoT network into a package less than 15 feet long. The car’s networks include at least 80 electronic control units, up to 100 sensors and more than 50 pounds of copper wiring that in luxury cars can consist of 1,500 copper wires totaling a mile long. The entire wiring harness weighs more than 100 pounds, the heaviest part of the vehicle after the engine and chassis.

It wasn’t always like this. The average family car in 1948 had about 55 wires totaling 150 feet long, but today’s vehicles are rapidly becoming an electrical/electronic tour de force, an IoT network on wheels.

In the coming all-electric automotive future, vehicles may be less mechanically complicated than their fossil-fuel-powered counterparts, but their wiring harnesses will actually account for a larger portion of overall vehicle weight. They also require expensive high-voltage lines for AC-DC charging and special wiring for the high-voltage battery and the battery management system.

When autonomous vehicles arrive, they will have even more sensors, all of which will be connected with wires and operate over a half dozen or more integrated networks with information shared across them all. Many will be operating at blazing speeds  — 10 Gb/s or more — that will place new demands on achieving low latency while operating at frequencies of 5 GHz, at which rejection of (electromagnetic interference) EMI and radio frequency interference (RFI) is a major challenge because as data rates increase, the signal-to-noise ratio decreases.

Staving off disaster

From an automakers’ perspective, this is a recipe for disaster because this level of complexity is no longer manageable. There are just too many electronic control units (ECUs), sensors and cables to handle, and they are not organized in an effective way. The only reasonable means of addressing this challenge is to move from the (controller area network flexible data-rate) CAN-FD bus to an Ethernet network, consolidate functions currently performed by multiple ECUs into a single device, and perform in software functions previously performed by hardware. Together they will not only enable a roadmap for the future but dramatically decrease cost by reducing the amount of heavy, expensive cabling.

Fortunately, the auto industry is taking steps to solve this problem, revolutionizing the way vehicles are built and managed using a zonal E/E architecture that builds on the current domain controller architecture. It optimizes electrical and electronic resources with a high-speed communications backbone and increased multi-functionality of electronic components. In essence, it is a near-complete departure from the way vehicles are designed today.

Graphic of a car illustrating in-vehicle networking

Built on the current domain controller architecture, a zonal E/E architecture optimizes electrical and electronic resources with a high-speed communications backbone and increased multi-functionality of electronic components.

To achieve this, the zonal architecture introduces two new device classes, the vehicle server and zonal gateway. The server is designed to reduce the number of ECUs and the zonal gateway acts as a connectivity hub that aggregates communications through a single Ethernet link to the backbone. To reduce the workload on the server, some processing will be performed in edge devices, such as sensors. From a processing standpoint, there are two domains in which computation is performed: application and real-time, the former handling primarily high-level functions and the latter dedicated to time and safety-critical tasks. In each domain, reliability and security are fundamental.

To solidify this approach in a standard, the Mobile Industry Processor Interface (MIPI) Alliance last year released MIPI A-PHY v1.0, the first industry-standard SerDes physical layer interface and the only standard interface that supports native camera (CSI-2) and display (DSI-2) interfaces for automotive applications. It provides an asymmetric data link in a point-to-point topology that enables high-speed unidirectional data, embedded bidirectional control data and power delivery over a single cable.

Built on the current domain controller architecture, a zonal E/E architecture optimizes electrical and electronic resources with a high-speed communications backbone and increased multi-functionality of electronic components.

The standard’s focus on high-speed links for cameras, displays and sensors through CSI-2 and DSI/DSI-2 interfaces ensures that the required support for autonomous driving will be available, along with high reliability, low packet error rate (PER), extremely high immunity to EMC, cable lengths up to 15 meters with four inline connectors, data rates up to 16 Gb/s that is scalable to at least 48 Gb/s and an uplink data rate to 200 Mb/s.

MIPI A-PHY solutions use pulse-amplitude modulation (PAM), a multi-level signaling scheme that achieves 16 Gb/s for four levels of PAM (PAM-4) signaling or 16 levels (PAM-16) to reduce the symbol rate and cable signal loss. This approach makes it possible to achieve just-in-time noise cancellation and physical-layer high-speed retransmission for error correction that guarantees that all data packets will be successfully passed across the link. In addition, multiple sensors can feed a single ECU or multiple SoCs for redundancy and increased processing power. To be “future-proof,” MIPI A-PHY can scale in the number of sensor modules that can be connected to the same ECU and can be used for ECU-to-ECU connectivity.

A single road ahead

The auto industry is moving rapidly forward from generating power from fossil fuels to electricity (and possibly hydrogen), but regardless of the propulsion technology, the number of sensors and other electronic components will continue to increase. All these functions translate into the need for vastly increased processing power and data transmission at speeds that dwarf what CAN-FD can deliver.

Taking a cue from the way industrial IoT has progressed, the solution is to increase the number of functions an ECU can perform, distribute real-time processing to the edge of the network where the data is generated, and orchestrate them all via a modern, high-speed communication bus — Ethernet. This is how vehicles will be built in the future because consumers want more features and autonomous vehicles will demand exceptional security and reliability. There is no other way.

About Avnet

With a century of innovation at our foundation, Avnet can guide you through the challenges of developing and delivering new technology at any — or every — stage of your journey. We have the expertise to support your innovation, turn your challenges into opportunities and build the right solutions for your success. Make your vision a reality and reach further with Avnet as your trusted global technology partner.

 

Avnet

This content is provided by our content partner Avnet, a global technology solutions provider with end-to-end ecosystem capabilities. Visit them online for more great content like this.

This article was originally published at Avnet. It was added to IoTplaybook or last modified on 08/20/2021.