Using 5G to Grease the Wheels of Edgier, More Intelligent Automobile Domains

Significant developments such as the rollout of multi-gigabit 5G cellular infrastructure, real-world deployment of artificial intelligence (AI) and machine learning (ML) and the deployment of edge/cloud computing have been taking place. In this article we revisit the purpose and functions of vehicle domains before discussing how they can benefit from the application of these burgeoning technologies, with a view to bringing CASE vehicles a step closer.

Connectivity Domain

The connectivity domain of an automobile covers all onboard and external vehicle communications. It manages wired and wireless interfaces that are used for communication between onboard modules and those that communicate with other vehicles and the internet. It is responsible for the secure transfer and exchange of data and information using internal and external interfaces and communication protocols. These include radio, cellular, Wi-Fi, Bluetooth Low Energy (BLE), GPS, V2X, LIN, CAN, and Ethernet. External interfaces, both wireless and cellular, will use sophisticated integrated intelligent antenna modules that will allow interfaces to be added or removed as required. A 5G radio promises be the most important part of this domain, providing the vehicle with a low latency (5ns), multi-gigabit (10Gb+) data connection to the cellular network and which will be a key enabler, supporting the function and operation of all other vehicle domains.

Perception and Driver Replacement Domains

The purpose of these domains is to provide a vehicle with the intelligence to sense and interpret its environment so that it can independently and safely control the task of driving. The sensing components that are required to make this possible include radar, multiple cameras, LiDAR, sonar and other components used for vehicle positioning and environmental telemetry. Path planning, situational assessment, safety algorithms and the fusion of sensor data and are just some of the intelligent features that must be released to allow a driver to abdicate full control to a vehicle and in this regard, AI and ML will be critical to allow a self-driving car to continuously learn about its environment. By their very nature, AI and ML require huge volumes of data to be collected and sent to the cloud for offline mining followed by algorithm development. The benefit of using 5G for this purpose can be quickly illustrated by the fact that the amount of time taken to transfer the data collected by a CASE vehicle in a week, using a standard Wi-Fi connection, is 230 days whereas with a 5G connection this is reduced to only 2 days. This makes it feasible for a self-driving car to be parked overnight and connect to the cloud to upload the data generated during the course of the day, while also downloading new algorithms and functions to provide it with added capabilities for use on the following day. However, while cloud connectivity and processing are key to algorithm development, until 5G coverage becomes ubiquitous their application in safety critical applications will require the use of edge/cloud technology that has the ability to process sensor data and react accordingly in real-time.

Powertrain and Vehicle Dynamics Domain

Vehicle motion is the responsibility of this domain and in self-driving cars, motion can be automatically adjusted and optimized to meet personal preferences and environmental factors, such as weather and road conditions. The powertrain has been a feature of automobiles since the earliest designs and its purpose is to convert the fuel source into power that is delivered to the road surface. The powertrain includes the engine, gearbox, drive shaft, axle, and wheels, all of which operate in harsh conditions that include high temperatures, strong vibrations and changeable weather conditions. The vehicle dynamics portion of this domain includes the steering and suspension which provide stability and a smooth driving experience and are the location for many of the sensor technologies that provide safety and comfort. In this domain, edge/cloud controllers can be deployed to provide real-time control in the event of an emergency maneuver (change of direction, speed reduction etc.) while AI/ML opens the possibility of better understanding how vehicle motion is impacted by changing road and weather conditions enabling safer and more energy efficient powertrain designs. The large volumes of telemetry information generated by powertrain sensors could also be transmitted using 5G traffic management infrastructure, allowing them to respond immediately in the event of vehicle breakdown and also to relay information about vehicle condition to the manufacturer who could send over-the-air updates to immediately fix performance anomalies that are detected.

Click image to enlarge

Figure 2. Various types of sensors for power train, body control, safety and information-related tasks support the electrification of the automobile and the spread of ADAS to improve safety and convenience

 

Body and Comfort In-Vehicle Experience Domains

The body and comfort domain entails the more basic functions provided for the driver and passengers and is where manual safety mechanisms (seatbelts) and access mechanisms (door locks) are managed. The domain has the ability to learn individual preferences from the behavior of vehicle occupants, including the settings that people prefer to have in place when they are in the car – seat and mirror position, climate control settings, temperature – which can all be automatically adjusted to personal taste. These functions, such as window controls and seat adjusters, use legacy auto-electronics and typically perform hardware operations using software for greater control and better adjustment. Sensors and microcontrollers help to create smart lighting functions that improve safety and respond to personal circumstances. For exterior lighting, headlights automatically adjust to weather conditions and the presence of oncoming traffic. Programmable interior lighting can be zoned to make it easier for passengers to sleep, read or watch a video, and dashboard settings can be adjusted automatically based on time of day and the occupant’s state, for example, sleeping. The in-vehicle experience domain relates to the productivity, entertainment and well-being of those onboard and aims to provide a similar experience to the driver and passengers to that which they would experience at home. It provides convenient access to digital content and the ability to create and customize new content. The software used in this domain must be both flexible and easily upgradable, to ensure access to content via existing hardware and also requires a sophisticated, touch-free human-machine interface (HMI) that can support voice commands, gestures and augmented reality. Key requirements for this domain experience include over-the-air (OTA) updates, the ability to monitor and learn, software upgradability/flexibility for content access and advanced HMI. High-speed 5G data transfer can be deployed in this domain to enable real-time downloading of entertainment (videos, games etc.) and to provide OTA software updates while AI and ML can be deployed to learn the behavior of and automatically customize the experience of vehicle occupants.

Click image to enlarge

Figure 3. The number of onboard electronic components continues to grow with the increasing functionality of automobiles, such as automated driving, and the associated introduction of more electrical components

 

Energy Domain

The energy domain covers the vehicle’s energy sources and storage units, energy harvesting and regeneration functions and also the power electronic modules, hardware, software, firmware and high-level algorithms that control the functions of batteries and supercapacitors. It is responsible for the overall management of charging, discharging and recharging of power cells. The battery management system (BMS) is one of the key components for achieving a safe and efficient energy storage and usage system based on batteries. It plays a critical role in reducing costs, accurately estimating the battery state and monitoring overall safety, (preventing overcharging and overcurrent conditions from occurring). In this respect, AI and ML techniques have been shown to provide better insight into battery state and performance than traditional modelling approaches and through continuous updating of algorithms via a high-speed cloud connection, offer the promise of improved energy optimization techniques in this domain. Given the power-hungry nature of AI and ML processing, they could even potentially learn ways to improve their own efficiency.

Conclusion

The domain-based vehicle architecture is much better suited to the development and development of CASE automobiles but each domain in turn has a specific set of requirements. In this article, we revisited the main functions of each domain in an automobile and showed how that can benefit enormously from more recent technologies including high-speed 5G cellular infrastructure, cloud/edge and AI/ML to assist with the development of smarter, safer and more energy efficient vehicles.

TDK Corporation