Autonomous driving is one of the megatrends shaping the automotive industry, driven by Software-Defined Vehicle (SDV) architectures. While shipments of fully self-driving cars will remain small throughout the rest of the decade, semi-autonomous (Level 2, Level 3) features are already becoming common in the smart mobility market. From active safety features many automakers are already familiar with to automated highway cruise control, autonomous driving technology is maturing. However, more advanced applications still need development time before being safe for mass deployment. This article will evaluate the primary applications, benefits, and enabling technologies dictating the future direction of autonomous driving.
Chart 1: New Vehicle Shipments by Level of Automation
(Source: ABI Research)
Benefits of Autonomous Driving
Driving is a repetitive and burdensome activity prone to many safety risks. Generally speaking, humans do not excel at performing driving tasks and are easily distracted. Therefore, automating or semi-automating the driving experience is a practical next step to improving automotive experiences. The list below outlines the three main benefits of autonomous driving:
- Safety (Short-Term Opportunity): As an SAE Level 1 feature, active safety systems detect road obstacles and automatically adjust steering or hit the brakes to avoid collisions. These use cases are already well-established in the automotive industry.
- Reclaimed Time (Mid to Long-Term Opportunity): While Level 2 driving autonomy requires human driver oversight, Level 3 and Level 4 allow drivers to disengage in many scenarios. This allows drivers to spend more time on non-driving tasks and activities.
- Sustainability (Mid to Long Term): In the future, consumers can order a driverless shared vehicle service for their commute. Although small-scale deployments exist worldwide, there are still some safety and functionality kinks to iron out before full-scale deployment is possible. Fleets of driverless taxis will lead to better use of shared assets, satisfying the demand for personal mobility with fewer vehicles. This will reduce congestion and free up dedicated parking areas for other urban use cases.
Autonomous Driving Applications
The market for Autonomous Vehicles (AVs) will unfold along two key dimensions: feature automation and driver disengagement. While the former includes basic active safety features, the latter is closer to the fully autonomous driving experience the industry envisions.
Active Safety and Driver Convenience
Thanks to the General Safety Regulation (GSR) 2 in Europe and safety rating agencies in the United States, automatic active safety is a popular application for self-driving. Advanced Driver-Assistance Systems (ADASs) are a staple in the automotive industry, with roughly 60% of vehicles shipped in 2023 equipped with them.
An ADAS solution provides the following safety features:
- Automatic Emergency Breaking (AEB)
- Lane Departure Warning/Lane Keeping Assist (LDW/LKA)
- Blind Spot Detection (BSD)
In addition to active safety features, ADASs also make the driving experience more convenient, providing Adaptive Cruise Control (ACC), Relative Speed-Based Routing (RSR), and High Beam Assist (HBA).
Level 2+/Level 2.5
Autonomous vehicles can have features like speed and following distance control, lane keeping, maneuver overtaking, exit taking, and traffic merging. Today’s Level 2 (L2) systems leave much to be desired, lacking many key features for the full autonomous driving experience. ABI Research observes that Level 2+ or Level 2.5 will inch closer to truly automated car functionality. These more advanced systems can automate enough driving tasks to support point-to-point navigation whereby every maneuver is automated on the driver’s behalf, whether in the city, in the suburbs, or on the highway. However, these AV applications still require driver supervision. Not until Level 4 of vehicle autonomy will complete eyes-free applications be achieved.
Automated Highway Driving
Highway driving involves repetitive tasks for long periods, making it a relatively easier AV application for Original Equipment Manufacturers (OEMs) to target than complex urban environments. This AV use case will be especially appealing for cross-country drivers. Not only does automated highway driving give drivers back more time, but it also ensures a safe trip when the driver gets sleepy behind the wheel.
Highway driving is easy to target for automation because the journey is predictable; there are no cyclists to worry about, no children running around the street, lanes are well marked, and everyone is traveling in the same direction. At the same time, automated driving addresses the boredom that drivers often experience on the highway, such as sitting in a traffic jam. The application allows drivers to perform more enjoyable tasks, such as reading a book during certain scenarios. In this context, automating highway driving is appealing in two ways: it’s relatively easy to implement for OEMs and it improves the driving experience. ABI Research sees this being be a short to mid-term opportunity for OEMs, providing excellent value for consumers.
Point-to-Point Navigation
While a longer-term opportunity, Level 4 vehicle autonomy will be the bridge to full, point-to-point navigation in the future. This smart mobility application primarily addresses the complexity of urban driving scenarios, which require more maneuvers and stops. Whereas Level 3 systems will still require the driver to be available as a backup, Level 4 automation safely drives the car without human intervention.
Robotaxis
Once full-fledged AV applications become safer, cost-effective, and mature, there will be less need for human taxi drivers (or Uber, Lyft, and other ride-sharing service drivers). Our analysts foresee the rollout of “robotaxis” as a gradual process with some bumps along the road. For example, the recent functionality issues of Cruise autonomous taxis in San Francisco highlight the immaturity of AVs in dense city environments. The path to AV-based shared mobility will also include adopting semi-autonomous features in passenger vehicles.
Further Reading
Imaging Radar’s Potential to Undercut LiDAR for Autonomous Driving (AN-5806)
Enabling Technologies for Autonomous Driving
Behind every autonomous driving application is technology, no matter how basic or advanced the application. Below is a list of some of the key technological underpinnings of AVs.
- Sensing: To automate driving perception, AVs use camera sensors to identify other vehicles, pedestrians, boundaries, and obstacles. Radar and Light Detection and Ranging (LiDAR) sensors are sometimes required to overcome visual barriers, such as poor weather, occlusions, and extreme lighting.
- Computing: Autonomous driving requires fast response times, which stipulates the deployment of local computing resources. The compute requirements vary greatly, with basic active safety applications requiring as little as 0.25 Tera Operations per Second (TOPS) and unsupervised automation applications requiring as much as 350 TOPS to 1,000 TOPS.
- Software and AI/ML: AV stacks are formed by scalable software modules that guide perception, sensor fusion, localization, path planning, and motion/control features. Artificial Intelligence (AI)/Machine Learning (ML) techniques are invaluable for AV developers that must account for the many complex driving scenarios in which an AV will find itself.
- Digital Maps: Accurate maps are essential for the autonomous vehicle to complete a route timely and safely. It’s important that the maps are updated in real time to reflect traffic and weather conditions. Intelligent Speed Assistance (ISA) and some other active safety systems would also benefit from digital maps.
Autonomous driving is just one of the three megatrends defining the future of smart mobility, alongside the software-defined car and electrification. To learn more about the advantages, technologies, and applications related to these automotive trends, download the free whitepaper: The Software-Driven Megatrends Shaping The Automotive Industry.
