Bench Talk

(Source: Masque / stock.adobe.com, generated by AI)

Automotive megatrends such as autonomous driving (AD) and advanced driver-assistance systems (ADAS) are challenging designers to convert from traditional electronic control units (ECUs)^[1] to domain control units (DCUs),^[2] and to further accelerate adoption of central communication units (CCUs) and zone control units (ZCU).^[3] According to the McKinsey Center for Future Mobility,^[4] the next generation of electric/electronic (E/E) architecture using zonal computing has started and expects a compound annual growth rate (CAGR) of around 44–45 percent between 2023 and 2030. The DCU usage will increase to US$44M value during the same timeline (Figure 1).

Figure 1: Automotive Compute Units Market ($Mio USD) (Source: McKinsey Center for Future Mobility)

The typical domain-based electrical architecture that dedicates a control unit for each functional system is giving way to a zonal approach;^[5] this evolution is fundamentally driven by four factors:

1. Facilitation of secure over-the-air (OTA) updates: Large numbers of ECUs can cause update bottlenecks and complexity, leading to safety, reliability, and regulatory compliance challenges. The consolidation offered by domain or zonal architecture is a significant step forward that will also facilitate rollbacks when updates fail.
2. Modular and collaborative hardware and software: Modular and collaborative hardware and software offer the potential to speed up development times. They are enabling faster reengineering and accelerating time to market.
3. Increasing silicon consolidation and integration: Zonal controllers achieve silicon consolidation and integration by adopting the functionality of several ECUs. Meanwhile, smaller node sizes boost power efficiency. Emerging “system-on-chip” (SoC) designs bring together several central processing unit (CPU), memory, and dedicated hardware (HW) accelerator subsystems. Unsurprisingly, this means that the latest SoCs for zonal controllers are based on node sizes of 16 nanometers (nm) and below.
4. Reduction of wiring harness complexity: Because zonal controllers serve as input/output (I/O) aggregators and are often placed at the mechanical construction of the car, they enable a less complex wiring harness, which in turn can foster standardization, support automation in the production process, and reduce costs due to lower employee skills needs. Wiring harness costs in a modern vehicle often account for 20 percent of the total E/E architecture budget, which amounts to a significant benefit.

In recent years, autonomous electrical vehicle development has gained significant attention from researchers and engineers.^[6] Many changes are revolutionizing automotive EE architecture, leading to autonomous driving and its associated challenges. An autonomous vehicle must be capable of sensing the environment and safely navigating without human input.^[7] The US National Highway Transportation Safety Administration (NHTSA) has defined five different levels of autonomy,^[8] ranging from no automation to full automation.

On the higher autonomy levels, the vehicles must sense their surroundings using multiple sensors, such as LiDAR, cameras, GPS, etc. Based on the sensor inputs, the vehicles must locate themselves in real time, make decisions, and act on driving. The sensors aim to improve driving safety and are called Sensing. On the other hand, the AD systems’ ability to process the data from sensing, make decisions, and command orders to actuators (brake, steering, etc.) is named Cognitive, as shown in Table 1 and Figure 2.

Table 1: The higher-order steps of information gathering and decision making in higher autonomy levels. (Source: Mouser Electronics)

Figure 2: Automotive placement strategy with polymer capacitors (Source: KEMET Electronics, an YAGEO Group Company)

The computing systems for AD can be divided into computation, communication, storage, security/privacy, and power management.^[9]

The continuous effort to increase the autonomy level significantly enhances computing system capabilities for AD. According to Liangkai et al.,^[9] today's “state-of-the-art” computing systems for AD include seven performance metrics, nine key technologies, and eleven open challenges.

As the level of autonomous driving increases, the number of sensors installed must increase accordingly to acquire data on the surrounding environment. As the number of sensors increases, the amount of data processed by SoC increases, and the power consumption of the main semiconductor device that performs data processing increases. This evolution increases power consumption to optimize the cognitive “brain” capability of perception and decision making.

Inside a DCU schematic exists:

1. a transceiver circuit that communicates with sensing ECUs,
2. an SoC that processes data and makes decisions,
3. various memories (DDR/Flash),
4. a microcontroller MCU that controls driving based on the information assessed by the SoC, and
5. DC/DC converter power supply circuits to operate the various functions with different voltage rails (<1V, 3.3V, 5V etc.).

The SoC with low voltage, lower 1V, and typically > 25A, on the DCUs/ZCUs/CCUs requires components capable of high current, low loss, miniaturization, high-frequency operation, and high accuracy (voltage).

It is common for a DC/DC converter to use T598 polymer capacitors for noise reduction at the input and smoothing/decoupling at the output. Typically, the T598D476M025ATE060 (EIA 7343-31 47uF25V, 60mOhm) has been adopted at the input and the T598D477M2R5ATE006 (EIA 7343-31 470uF2,5V, 6mOhm) has been successfully designed at output smoothing/decoupling. To further advance future needs, KEMET now has prototype samples available for the next generation of input noise reduction, with the T598D107M025ATE050 (EIA 7343-31 100uF25V, 50mOhm) for capacitance extension and the T598D687M2R5ATE006 (EIA 7343-31 680uF2,5V, 6mOhm) for optimum output smoothing.

The demand increase in power consumption to support higher processing data and action will continue to require the T598 series' main advantages: high capacitance combined with low capacitance roll-off in frequency and temperature stability, low ESR and high ripple performance, and extended life span performance.

Conclusion

As the automotive industry continues to evolve towards higher levels of autonomy, the reliability and performance of tantalum polymer capacitors like the T598 series will be crucial in ensuring the safety, efficiency, and longevity of these sophisticated systems. Their low ESR, high capacitance, and exceptional stability across a wide range of temperatures and frequencies make them ideal for supporting the increased power consumption and data processing needs of advanced autonomous driving systems.

The "Software-Defined Vehicles ‘State-of-the-Art’ and Challenges with AD and ADAS Computing Systems" blog was authored by Cristina Mota-Caetano for KEMET Electronics, a YAGEO Group Company, and is repurposed here with permission.

Author

Cristina Mota-Caetano
Director Technical Product Marketing - Product Management – Tantalum BU
KEMET Electronics, an YAGEO Group Company
Cristina Mota-Caetano brings 27 years of experience in technology and product marketing activities on tantalum capacitors and acts as global technical product marketing responsible on tantalum business unit. She brings material science education background and experience from research and development to portfolio management and placement and demand creation initiatives.

Sources

[1] https://en.wikipedia.org/wiki/Electronic_control_unit
[2] https://autotech.news/autonomous-driving-and-cockpit-domain-control-unit/
[3] https://www.continental-automotive.com/en/solutions/server-zone-architecture/zone-control-units.html
[4] https://www.mckinsey.com/industries/semiconductors/our-insights/getting-ready-for-next-generation-ee-architecture-with-zonal-compute#/
[5] https://www.eetasia.com/the-role-of-centralized-storage-in-the-emerging-zonal-automotive-architecture/
[6] Overview analysis of recent development of Self-Driving Electric Vehicles,” Qasim Ajao and Landre Saqeeq, Georgia Southern University.
[7] “CAAD: Computer Architecture for Autonomous Driving,” Shaoshan Liu, Jie Tang, Zhe Zhang, and Jean-Luc Gaudiot, IEEE.
[8] Policy of Automated Vehicles, NHTSA
[9] “Computing Systems for Autonomous Driving: State-of-the-art and Challenges,” Liangkai Liu et al.