There have been some bumps in the road to autonomous vehicle adoption. From limitations associated with 4G LTE to spectrum reassignment, it has not been as smooth a ride as many had hoped. Technological advances, the continued deployment of 5G, and evolving architectures are helping to create a fast track to autonomous vehicle technology gaining traction. They are also requiring new test processes.
The Society of Automotive Engineers (SAE) has spent considerable resources in defining standards for autonomous vehicles. Figure 1 outlines the J3016 Levels of Automated Driving that reflects the evolving standard.
Figure 1: The SAE J3016 Levels of Automated Driving
As you can see, six levels of driving automation, from SAE Level Zero (minimal automation) to SAE Level 5 (full vehicle autonomy) are outlined in the standard. It is widely recognized as the universal reference for autonomous vehicle capabilities. Market leaders, including the American Automobile Association and Transportation Research Board, had input in developing the levels.
Updates have been made to the standards, as technology has advanced. Due to trends in wireless connectivity for vehicles, driving automation systems have evolved from basic safety telematics to massive communications. It’s continuing, as 5G is helping in the rollout of next-generation architectures.
The Spectrum Impact
In May 2021, the Federal Communications Commission (FCC) issued its final rule to split the 5.9 GHz band, which had been exclusively for Intelligent Transport Systems (ITS). It now also includes unlicensed uses, especially increased Wi-Fi demands.
This decision came despite the objections of the Department of Transportation (DoT), which had exclusive control of the entire spectrum. Despite the re-allocation of the ITS spectrum, autonomous automobile systems still will be able to operate highly efficiently through other methods, such as radar and LIDAR.
Those other methods benefit greatly from 5G, which delivers ultra-reliability, low latency, and high throughput. A further indication of the positive impact of 5G can be seen in 3GPP Release 16. The standard includes Cellular vehicle-to-everything (C-V2X) Sidelinking (SL) with 5G New Radio (NR). Cellular networks are no longer necessary to enhance semi-autonomous and autonomous driving because of SL, allowing for more advanced autonomous driving use cases.
Advanced Driver Assistance Systems (ADAS)
A cornerstone of autonomous vehicles is ADAS. In some respects, it has served as a bridge connecting legacy vehicles to full self-driving automobiles. In fact, nearly 93% of all new cars as far back as 2018 had at least one ADAS feature, helping to set the stage for future smart car designs that are leading toward autonomous vehicles becoming mainstream.
ADAS established the first driver-assistance systems, introducing advanced electronics and wireless technology into the automotive world. From warnings notifying drivers of forward collisions and lane departures to adaptive cruise control and anti-lock brakes, ADAS is making driving safer. As many as 2.7 million collisions can be avoided annually in the United States, because of ADAS.
Of course, the widespread adoption of ADAS and autonomous vehicles is based on the successful operation of the telematics control unit (TCU). These onboard embedded systems connect the various elements of smart vehicles. Autonomous vehicles need reliable communication to download the ITS-related dynamic information and to upload vehicle sensor information as service data through the TCU.
For these reasons, connectivity tests on TCUs and the systems within the vehicles is imperative. It is required to test for robust communication. Table 1 outlines the key verifications in 5G TCU development.
Table 1: Necessary tests to verify TCU performance
Anritsu, often collaborating with similar market leaders such as DSpace and Spirent, develops solutions to emulate autonomous vehicle design use cases to verify connectivity. The Anritsu-based solutions simulate a 3GPP network (with network slicing), connecting the application servers to the car core computer via 5G and multi-access edge computing (MEC). It can perform scenario-based tests and parametric tests, load tests, reliability, and regression tests on a single connected platform
Evolving to Zonal Architecture
Another evolution engineers must consider involves functional architectures. There is a shift from the traditional distributed architecture to the current domain centralized architecture. Zonal architecture, which groups vehicle systems by function, is an approach that better supports ever-more complex ADAS and autonomous driving systems. These architectures are also changing how testing needs to be conducted.
In a domain architecture, modules that provide similar functions and related controlled devices or sensors are connected via electronic control units (ECUs) to dedicated domain controller units (DCUs). This reduces the amount of ECUs and cabling, however, reliability becomes more challenging as the sensors and devices become farther from the ECU.
A zonal architecture is more efficient, as it’s a cluster of high-performance computers creating a network system that spans the vehicle. Cabling connects the central processing module to the zonal gateways and sensors. Automakers can easily scale zonal architecture to support additional sensors and electronics throughout the vehicle.
Verifying connectivity within a zonal architecture requires a specific testing approach (figure 2). It varies significantly from connecting testing for the domain architecture (figure 3).
Figure 2: Test configuration for verifying zonal architecture
Figure 3: Test methods for verifying domain architecture
To learn more about conducting tests on automotive system designs, you can view this webinar.