Electrified or not, autonomous or not — today’s modern vehicle needs energy for more than just getting to its destination. On-board intelligence is now found in even modestly priced cars, and all these smart systems must draw power from somewhere.
With advanced driver-assistance systems (ADAS), diagnostics powered by software, and entertainment systems now standard on most vehicles, today’s modern vehicle is often described as a “server on wheels.” It could also be described as a cluster of endpoints, full of memory and storage, sensors, processors, and network connectivity sharing data internally but also outside the vehicle. The more autonomous the vehicle, the more on-board intelligence is present, and the more components required to make all the various smart elements work in concert quickly and reliably.
All these components must meet certain requirements as standalone devices, from a reliability and power consumption perspective, while also being able to work together as a whole without competing for resources in a manner that would compromise the performance and functional safety of the vehicle.
More autonomy means more components that draw power
Most new vehicles today come equipped with ADAS, and it’s important to remember that there’s an “s” on the end of “systems.” It’s not just one device, but a collection of devices.
Unless you’re still driving an old vehicle, your car has some form of ADAS onboard, which is aligned with the six levels of driving automation. Level 0 means exactly that — everything is done manually. Level 1 provides some driver assistance, usually in the form of cruise control. At Level 2, there’s partial automation of steering and acceleration, but the driver is still responsible for monitoring the environment. Level 3 and up is where increasing automation kicks in and the automated system monitors the environment. We’re only just getting to Level 3, with Levels 4 and 5 still relatively far ahead on the horizon.
For most drivers, ADAS in the car is expressed through beeps as you back up to let you know you’re getting closer to an object, along with a rear-view camera so you don’t have to strain your neck to look behind you, while cruise control is now a standard feature. These features require computing power and memory as well as connectivity to get data from the camera to the screen. More importantly, they must be responsive and start up quickly. The majority of ADAS components are passive and contribute to overall reliability. They include self-healing capacitors that support fail-safe power supplies and regulators in automobile circuits.
The smarter the feature, the more active the components, encompassing cameras, sensors, and various displays that depend on some form of compute, memory, and data storage, which, in turn, consume power as well as generate heat. This requires balancing power consumption and managing thermal dissipation.
Moving beyond blinking lights
Not all the smarts in today’s vehicles are just for the driver. The “check engine” light has evolved into on-board diagnostic (OBD) systems, which, like ADAS, are also standard on a wide range of vehicles, including the average family car or commercial trucks. Behind the lights on the dashboard are increasingly complex systems that are constantly monitoring the health of a vehicle and even logging data. Today’s mechanics are just as likely to plug a computer into a vehicle to diagnose an issue as they are to poke their head under the hood. OBD systems are now a key aspect of vehicle maintenance and repair.
Just like ADAS, OBD systems require a whole host of components and electronics — all of which draw power. At the center of any OBD system is the electronic control unit (ECU), which gathers data from various sensors throughout the vehicle and uses it to control parts of the vehicle such as fuel injectors or monitor for issues.
And just as ADAS employs sensors to monitor the environment of the vehicle, the OBD system uses sensors through the vehicle to monitor the electronics, chassis, and engine. Not all data is saved, of course, because that would require too much storage capacity. Instead, the system retains information only when the input falls outside of the normal range in the form of a diagnostic trouble code (DTC). A DTC will lead the ECU to send a signal that turns on an indicator light to you know there’s a problem — a smarter check engine, if you will. Today’s indicators are nuanced in that you can tell whether it’s an urgent problem or something that can wait.
Beyond vehicle diagnostics, OBD systems can aid with emissions testing and support commercial vehicle telematics for entire fleets, gathering information about fuel efficiency, driver behavior, and remote diagnostics to guide preventative maintenance schedules.
A fun ride requires more energy
ADAS and OBD systems are just one half of the “infotainment” that’s found in most vehicles today. Entertainment systems have gone beyond spending a little more for a CD player rather than just a radio. In addition to on-board GPS, more cars include satellite radio as well as the connectivity required for roadside assistance services such as On-Star or passenger internet, whether you choose to subscribe to it or not. Larger family vehicles such as minivans may even have screens to display video content to keep young passengers entertained on long trips, and your car may be able to store music and videos rather than relying on you to plug in your phone to access your collection.
All these systems are why the “server on wheels” analogy is less apt than calling the modern vehicle a cluster of endpoints. And endpoints, of course, need to be connected, whether it’s a memory storage device, a camera, or a diagnostic signal. This connectivity moves the data back and forth, whether it’s within the vehicle or to the external receiver, and this also requires power.
Connectivity drives power consumption
For the links to connect various ADAS and OBD system elements, radio frequency (RF) capacitors and inductors are increasingly being replaced by electrostatic-discharge–capable inductors, as they can harden RF links and sensors throughout ADAS designs, while pulse-capable integration capacitors can be used on IR cameras, stereo vision, and lane-keeping modules to improve system reliability.
The major evolution of vehicle connectivity, however, is 5G, which is a major contributor to smarter vehicles, especially the autonomous functionality of the future. It’s now enabling access to vehicle telematics data to support fleet management and maintenance. The more data that can be collected easily and quickly, the more insight is possible, which benefits those designing automotive systems, fleet owners, and drivers.
Connectivity also enhances ADAS because data gathering by the vehicle can be complemented by external sources, including the cloud and other vehicles. The faster speeds of 5G will also be required to fully realize autonomous driving as well as provide necessary security to share AI inference data and software updates — in real time, if necessary. The entertainment features for passengers will also take advantage of 5G connectivity to access subscription content using the same networking equipment as more essential data services without compromising performance, reliability, or security.
Components must balance power and performance
Whether it’s for compute, memory, storage, or the connectivity required to move data, power consumption must be a consideration. Processors that are needed for data-intensive tasks in the car, such as computer vision and deep learning, will need to handle streams of information and process it in real time while operating within specific power budgets. Lower-power processors can also reduce or even remove the need for active cooling.
The adoption of ADAS Level 2 and Level 3 to support features such as adaptive cruise control, lane keeping, automatic braking, and driver-monitoring systems increases demand for memory content in vehicles as well. For small amounts of data that must be stored throughout devices in a vehicle, established memories such as NOR flash offer fast-boot capabilities, so functions come to life as soon as the driver turns the key without drawing power when not needed.
For functions that must be processed quickly but require higher-capacity memory, low-power, energy-efficient memory such as LPDDR4X/5X may offer the best balance between power and performance. More advanced process modes will be essential to realize Level 4 and Level 5 autonomous vehicles that run AI-enabled applications that require both, although GDDR6 may be the preferred memory for some applications to support multiple screens in vehicles.
Consolidation contributes to better power profiles
Because there are so many devices running within a vehicle, architectures are at risk of becoming more fragmented, so there’s a move to having one domain controller for all ECUs, while automotive designers will take advantage of multi-chip packages to store memory into a single, streamlined package. Another approach to consolidating systems is to combine data storage for critical systems with non-critical systems, with enough smarts built in to prioritize operations.
That’s where NAND flash comes into play, but rather than having multiple, discreet eMMCs or UFS flash devices, storage will be centralized and consolidated by taking advantage of SSDs, which could hold both information essential for driving and entertainment content. Infotainment systems with maps are getting higher in resolution, and that will further drive the need for an architecture that combines computing and memory together, as well as expand needed flash capacities, especially as more and more data is logged.
Emerging memories that are lower in power consumption also show promise for automotive applications. Embedded MRAM has been demonstrated to be able to handle the extreme environment of automotive, while FRAM is well-suited for the high-speed non-volatile data logging needed for autonomous vehicles.
Ultimately, the power consumption profiles of processors, memory, and storage are a key factor in the development of smarter vehicles, especially as they are increasingly electrified. Just as fuel efficiency is valued in a car with a combustion engine, energy efficiency is becoming an increasingly important metric for smarter vehicles, regardless of their power source.
