Why is a car a data center on wheels? How to ensure the strong performance of the in-vehicle network?

“Comparing the past, present, and future of cars, there is a clear trend: cars have become data centers on wheels. Inside every car, the flow of data from safety systems, on-board sensors, navigation systems, and more, and the reliance on that data, is growing rapidly.

“

Comparing the past, present, and future of cars, there is a clear trend: cars have become data centers on wheels. Inside every car, the flow of data from safety systems, on-board sensors, navigation systems, and more, and the reliance on that data, is growing rapidly.

This has significant implications for In-Vehicle Networks (IVNs) in terms of speed, capacity, reliability, one of which is in high-speed low-latency applications such as Control Area Network (CAN), FlexRay, Local Interconnect Network (LIN) ), Media Oriented System Transport (MOST) and Single Side Half Wave Transport (SENT) buses lack the required bandwidth. As a result, these traditional standards are gradually being replaced by previously proven technologies in the field of information technology (IT).

The current prime example is Automotive Ethernet, which covers four standards developed by Electrical and Electronics Engineers (IEEE). Currently, Automotive Ethernet will coexist with a variety of buses covering various systems and subsystems. Therefore, we need different test methods to complete the design, validation, commissioning, troubleshooting, maintenance and servicing of automobiles and IVNs.

Trends: More demand comes from data, Ethernet, standardization and lifecycle.

Today’s cars have more than eighty Electronic Control Units (ECUs), CAN, LIN, FlexRay, MOST and SENT have always carried information between these ECUs and various onboard systems such as engine, drivetrain, transmission, brakes, Body, suspension, infotainment, etc. (Figure 1). In addition, cellular and non-cellular wireless technologies are streaming external data to infotainment, navigation and traffic information systems.

Figure 1: In automotive systems, different buses and data rates provide the necessary communication capabilities.

Process more data

In the next few years, we expect to see more than 100 ECUs in every car, with connected in-vehicle networks carrying several terabytes of data per day. We expect automotive to continue to adopt CAN, CAN-FD, LIN, FlexRay, SENT and MOST; however, the current top figures are 10 Mbps for FlexRay and 150 Mbps for MOST. Of course, “going faster” is always easier said than done, and the industry’s widespread adoption of the CAN bus requires a massive redesign to provide the necessary speed, security, and backward compatibility.

As sensors grow in number and sensitivity, they generate enormous amounts of data. It is conceivable that 10~20 cameras provide a 360-degree panoramic view, and all cameras send 1080p (now) or 4K (future) high-definition data streams, and the pixel depth is increased from 16-bit to 20-bit or even 24-bit. The numbers are quickly adding up: a 4K camera that supports 24-bit pixel depth produces 199 Mb of data per frame at 10-30 frames per second. While 1 Gbps rates may be sufficient now, 10 Gbps will soon be required (Figure 2).

Figure 2: Faster data rates and wider bandwidths are required between an increasing number of sensors and ECUs.

Currently, IVNs employ preprocessing hardware to perform data reduction (ie, compression) on the sensor. Unfortunately, this introduces latency, affects response time, and degrades image quality, limiting the available detection distance. An emerging solution is to deliver raw data at 2 – 8 Gbps to centralized systems-on-chips (SoCs) or general-purpose processing units (GPUs), which can compress incoming real-time data. IVNs are moving from a flat structure to a domain controller structure, where sensors send raw data to a central processing unit.

The required communication traffic is constantly expanding and evolving with vehicle-to-infrastructure (V2I), vehicle-to-vehicle (V2V), and vehicle-to-everything (V2X) technologies. All of these will play an important role in car operation and human-machine interaction.

Moving to Automotive Ethernet

In automotive applications, optimizing data utilization requires faster throughput, lower latency, higher reliability, and higher quality of service (QoS) to ensure safe and reliable vehicle operation. Automotive Ethernet will play an increasing role in carrying high-speed data communications as speeds reach 10 Gbps, including: IEEE 802.3cg, 10BASE-T1, 10 Mbps; IEEE 802.3bw, 100BASE-T1, 100 Mbps; IEEE 802.3bp, 1000BASE-T1, 1 Gbps; and IEEE 802.3ch, 10GBASE-T1, 2.5/5/10 Gbps.

Given the increasing data rates available and the demand for these capabilities, along with the need to reduce cable weight, many industry observers are optimistic about the growth of automotive Ethernet and the number of connected in-vehicle nodes.

Standardization: Gaining new business advantages

In the entire history of the development of the automobile industry, there is an established practice that has not changed for a long time, that is, standardization. This approach will still apply because standardization provides many important advantages, such as increased competition among manufacturers, lower component costs, and guaranteed interoperability.

Bus Topology and Data Rate Comparison

When considering different buses, the maximum data rate corresponding to each bus and the type of network topology it supports, Figure 3 summarizes some of the data. Automotive Ethernet also adds “switched fabric” capabilities to enable efficient performance in local area networks (LANs). It uses a combination of hardware and software, using multiple Ethernet switches, to control traffic in network nodes. A fabric network identifies all its paths, nodes, requirements and resources. Within this architecture, the available address space is 224 and up to 16 million nodes or devices can be connected.

In the field of next-generation IVN, examples of standardization include Automotive Ethernet, MIPI A-PHY and HDBaseT Automotive. By adopting proven technologies in the IT field, the automotive industry will gain significant new business advantages, as the car of the future is becoming a data center on wheels.

Figure 3: Major automotive buses are well suited for distance-specific tasks, but are therefore less versatile than Ethernet-based networks.

Life cycle: from development to maintenance

As cars become more autonomous, so do the consequences of system failures. To help ensure these systems operate safely and reliably, in-vehicle network testing continues to gain in importance throughout the vehicle lifecycle (Figure 4). Therefore, careful selection of system design tools and IVN test solutions to meet the needs of all different stages of the automotive life cycle will bring deep advantages to Tier 1 suppliers, automotive OEMs and automotive end users.

Figure 4: In-vehicle network testing is growing in importance throughout the vehicle lifecycle.

Multidimensional Challenge: Testing Multiple Buses Side by Side

Today, automobiles are employing various communication buses that operate simultaneously. Therefore, system optimization and debugging become very difficult and time-consuming. Using all of these technologies in parallel, and within the limited space of a vehicle, can lead to electromagnetic interference (EMI), poor signal quality, and possibly failure of critical systems.

Testing in-vehicle networks requires reliability checks throughout the vehicle, including interoperability, immunity to interference, crosstalk, and sources of interference. Verification of operational functionality and communication reliability will cover systems connected to the bus managed by each ECU within the vehicle (Figure 5). As vehicles become increasingly data-intensive, testing becomes critical to ensure safe and reliable operation at all stages of the lifecycle, including development, validation, production, maintenance and servicing.

Figure 5: The use of Ethernet as the central hub for communication between systems that currently rely on various dedicated buses.

Test Challenge #1: Debug Bus Issues

CAN, LIN, and FlexRay are relatively mature bus protocols designed for robustness and ease of integration. Even so, in-vehicle communications can be affected by noise, board layout, and startup/shutdown timing, resulting in excessive bus errors and lockups. Frequently asked questions for CAN, LIN and FlexRay include troubleshooting signal issues, debugging decoded protocols, understanding multiple channels, sensors and actuators. In SENT, it is difficult to configure the oscilloscope to decode fast and slow channel SENT messages before triggering on the decoded messages.

As mentioned earlier, multiple buses running simultaneously in the enclosed space of an automobile can generate EMI, resulting in poor signal quality. Pre-compliance testing can help you isolate and identify the cause of signal quality issues and bus performance issues, plus improve your ability to pass formal EMI and Electromagnetic Compatibility (EMC) testing against relevant standards such as CISPR 12, CISPR 25, EN 55013 , EN 55022 (replaced by EN 55032) and CFR Title 47, Part 15.

Test Challenge #2: Verify Electrical Conformance

Ensuring reliable low-latency data flow between and within vehicles is critical to the safe operation of the entire system. Unlike CAN, LIN, etc., Automotive Ethernet has a complex set of compliance tests specified by the IEEE and OPEN consortia, including various electrical requirements, to ensure compliance with the standard. These tests are typically performed during design, verification, and production. In automotive Ethernet, physical (PHY) layer electrical testing covers several key metrics of transmitter/receiver (transceiver) performance. The specific goal of these measurements is to test the consistency of the physical media attachment (PMA) against various electrical parameter data.

Test Challenge #3: Verify Protocol Consistency and System Performance

Automotive Ethernet operates at full duplex, so two linked devices can send and receive data at the same time. This provides three related advantages over traditional shared networks: first, two devices can send and receive data at once instead of taking turns sending and receiving data; second, the overall bandwidth of the system is greater; third, full duplex can Implement multiple sessions simultaneously between different pairs of devices such as master and slave.

In addition to these complexities, automotive engineers face another challenge: using PAM3 signaling for full-duplex communication makes it difficult to view automotive Ethernet traffic before fully characterizing signal integrity. To perform signal integrity analysis on a link and decode the protocol in a real system environment (using an oscilloscope), the designer must view each link separately, which requires signal isolation before analysis can be performed.

Test Challenge #4: Get the information you need to troubleshoot and debug

Whether the issue is bus performance, EMI, electrical conformance, or protocol conformance, there are two fundamental metrics that determine signal quality and, in turn, data performance, amplitude and timing. The accurate operation of these two indicators is essential to ensure the successful transmission of digital information through the bus. This is also becoming more and more difficult due to faster bus speeds and more complex signal modulation techniques (such as PAM3). Oscilloscopes are the measurement tool of choice, but without adequate frequency coverage, channel count, accessories, and on-screen analysis capabilities, troubleshooting and debugging can become tedious and time-consuming.

Solution: Harness the power of standardization

As mentioned earlier, standardization is a long-established practice in the automotive industry. Drawing from history, this concept can also be used to select solutions for IVN testing. Standardization can help you manage testing costs by adopting a uniform testing approach. For example, choosing a test platform that can easily accommodate higher speeds can allow you to increase the capital utilization of your test and measurement solution.

In the real world, we need to consider the different segregation of responsibilities throughout the life cycle of the vehicle and the onboard systems. Without a coherent strategy, common practices lead to random use of test hardware and software across different departments, and eventually build up. Let’s take a closer look at the general and specific features of the solution that can help you reduce testing costs while ensuring consistent results throughout the life of the vehicle.