Level 3 Autonomous Driving Explained: What It Means for the Automotive Industry

Level 3 autonomous driving represents a pivotal moment in the evolution of self-driving technology because it enables drivers to take their eyes off the road and their minds off the driving task.

In Level 3-Automated Mode, the Autonomous Driving System (ADS) monitors the road and drives the vehicle within its Operating Design Domain (ODD, e.g. highway driving or traffic jam). However, the driver is still the “fall-back” and must be “available” to take over vehicle control.

Level 3 autonomy is poised to transform how we drive and interact with vehicles allowing drivers to legally perform side tasks like answering emails or browsing to increase productivity and convenience

However, the transition of vehicle control poses safety challenges. Moreover, automotive players strive to find the perfect balance between system and component cost and robust performance. Finally, they must navigate the challenging regulatory framework requiring certifications and approvals which vary across major markets.

This article explores what Level 3 autonomous driving means, how it differs from other levels of autonomy, and its implications for automakers, suppliers, and consumers.

What is Level 3 Autonomous Driving?

Level 3 autonomous driving, as defined by the Society of Automotive Engineers (SAE), refers to a system where the vehicle can manage all aspects of driving in specific conditions without driver intervention. However, unlike Level 4 or Level 5 systems, the driver must be ready to take over control when requested by the system.

For instance, a Level 3-enabled vehicle can handle tasks like steering, acceleration, braking, and navigation in controlled environments, such as highways, but the driver must remain alert to intervene during unexpected situations.

Key Technological Building Blocks of Level 3 Autonomy

Operating Design Domain (ODD) for Conditional Automation: Unlike Level 2 systems that require constant driver supervision, Level 3 can operate independently in predefined conditions.

Carmakers are working on Level 3 features for:

• Cruising, such as Traffic Jam Pilot or Highway Pilot.
• Parking: summon, learnt-parking manoeuvres
• Understanding what these systems do and their limitations is crucial to avoid misuse and abuse.

Advanced Sensor Suite for robust performance: Level 3 vehicles rely on a combination of cameras, radar, lidar, high-definition mapping, AD-Domain Controllers, advanced AD-Software Stack and centralised E/E architecture.

There are different approaches to software and hardware stack, with some players working on vision-based autonomy, lidar-less and map-less Level 3.

Human-Machine Interface: The system notifies drivers when human intervention is necessary through visual, audio, or tactile alerts. Intuitive interaction with HMI and safe transition of control between the driver and the system are critical for higher levels of automated driving. The mandatory installation of driver-facing cameras and data storage aims at mitigating safety risks of ‘’transition demand’’

Over-the-air updates (OTA): system updates to enhance performance and fix issues are now delivered with SOTA-FOTA updates. Tesla gives the option to their customers to purchase ADAS after they purchase their vehicle (Feature on Demand) with OTA of software since the hardware is already installed.

Data Storage: The regulations also require vehicles to be equipped with a Data Storage System for Automated Driving — the so-called “black box” — which will record when ALKS is activated. This will help towards the determination of correct system operation as well as liability determination in case of an accident in “automated mode”.

Cyber Security and Functional Safety Requirements: ALKS will also need to comply with cyber-security and software update requirements set out in two new U.N. regulations (2021). Cyber Security becomes a regulatory requirement to tackle cyber threats and vulnerabilities. OTA updates will also play a crucial role to mitigate and patch threats.

How Level 3 Autonomous Driving Differs from Other Levels

Level 2 vs. Level 3

Level 2 systems provide partial automation, assisting with tasks like lane keeping or adaptive cruise control. However, the driver must actively supervise and remain in control. Level 3, by contrast, takes full responsibility under specific conditions, making it a significant leap forward.

Level 3 vs. Level 4

While Level 3 requires driver intervention in certain scenarios, Level 4 autonomy is designed to operate without human input in geofenced areas. Level 4 systems have higher reliability and can handle more complex scenarios autonomously.

Regulation for Level 3

The update of the UNECE R.79-Automated Steering to Regulation No.157- Automated Lane Keeping Systems (ALKS) unlocked Level 3-Driving features with speeds up to 60 km/h on motorways. The amendment was announced in mid-2020 and came into force in Jan’21 in Europe, Japan and China, among other signatories. In Jun’22, speed limits rose to 130 km/h.

Carmaker Roadmaps to Level 3 Autonomous Driving

Most carmakers follow an Incremental approach of adding longitudinal and lateral support to automate functions. For example, in Driving features, from L1-Cruise Control to Level 2-ACC to Cruise Assist, to Level 3-Highway Pilot.

A limited number of players aims at skipping L3. Collaborative business models to build L4 platforms and share costs and expertise are some of the key strategies for higher autonomy.

Carmakers are racing to commercialize Level 3 systems. For example:

• Audi: Audi showcased Level 3 capabilities in its A8 sedan but faced regulatory and infrastructure challenges that delayed wider deployment¹.
• In April 2022, BMW announced that its 7-Series to be equipped with a Level 3 autonomous system in limited scenarios on highways and moderate traffic at speeds up to 60 km/h. The BMW-7 Series got certification for Level 2+L3 in Germany
• Honda: In 2021, Honda introduced the Legend, equipped with the “Traffic Jam Pilot,” the world’s first certified Level 3 system, available in Japan². Honda’s Legend with L3 system called “Sensing Elite” went for sale in Mar’21.
• Mercedes-Benz: The 2023 S-Class and EQS models have received regulatory approval for Level 3 driving in Germany. This “Drive Pilot” system enables hands-free driving at speeds up to 60 km/h on designated highways³. The speed limit of Drive PILOT will increase to 95 km/h by 2025 (Sep’24).
• Nissan Skyline featured the ProPILOT 2.0 and the NissanConnect system introduced in 2020.
• Tesla has technical capability of L3+ in the US with the FSD, where regulations are different compared to Europe. L3 FSD is expected in 2025 in Europe and China pending approval (Sep’24)

Fragmented deployment of Level 3

Honda and Nissan launched Lv.3 in Japan only in 2021 leveraging the update of the UNECE regulation. BMW and Mercedes-Benz are launching Level 3 in 2024 in Germany & US. Changan claims they have Level 3 in China.

Lidar fitment strategies vary

In contrast to Tesla who does not feature a Lidar, BMW (iX & 7-Series) and Mercedes-Benz (L3-Drive Pilot) are among the carmakers equipping their new cars with Lidar for Lv.3+/4. They will offer L2+ until data validation and regulation allow Lv.3 via OTA.

Benefits of Level 3 Autonomous Driving

Enhanced Safety

Level 3 systems are designed to reduce human error, which is a leading cause of traffic accidents. By automating critical driving functions, these systems can improve reaction times and situational awareness.

Improved Driver Comfort

Drivers can delegate monotonous tasks like navigating traffic jams, allowing them to focus on other activities while the system takes control.

Industry Growth

The adoption of Level 3 technology is creating new opportunities in the automotive sector. According to a report by Auto2x, the global autonomous vehicle market is expected to reach $35 billion in 2025⁴.

Challenges and Considerations

Despite its promise, Level 3 autonomous driving faces several hurdles:

Regulatory Barriers

The rollout of Level 3 systems depends heavily on regional regulations. Some countries have yet to establish clear guidelines for conditional automation.

Infrastructure Limitations

Level 3 vehicles require high-quality road infrastructure, such as well-maintained lane markings and reliable GPS signals.

Safe Transition of Vehicle Control Between the System and the Driver

A significant concern is the “handoff” period—ensuring drivers can safely regain control when the system requests intervention.

The Future of Level 3 Autonomous Driving

The path to widespread adoption of Level 3 autonomy will be shaped by technological advancements, regulatory alignment, and consumer acceptance. With major automakers investing heavily in this space, the industry is moving closer to a future where vehicles can handle complex driving scenarios with minimal human input.

Artificial Intelligence to unlock Level 3 Autonomous Driving

AI plays a pivotal role in enabling Level 3 autonomous driving by enhancing a vehicle’s ability to make real-time decisions, interpret complex environments, and navigate with minimal human intervention.

In 2025, Honda will launch the Honda Series 0 with L3 Autonomy and Unsupervised Machine Learning from Helm.ai.

Honda’s Series 0 coming in 2025, equipped with Level 3 autonomy, showcases how advanced AI technologies, particularly unsupervised machine learning (UML), are transforming autonomous driving capabilities.

Unlike traditional supervised machine learning, which requires labeled datasets for training, Helm.ai’s unsupervised approach enables the AI to learn from unstructured and unlabeled data. This dramatically reduces the time and cost of training.

Demand for Sensors and Sofware to boost Suppliers’ Revenues

For Level 3 autonomy, a combination of high-performance sensors will be in high demand due to their critical roles in object detection, environment mapping, and decision-making.

The sensors that will dominate Level 3 autonomy are those that offer high precision, redundancy, and robustness across varying driving scenarios. LiDAR, radar, cameras, and driver monitoring systems will be especially in demand, supported by V2X communication and GNSS systems. With advancements in AI and sensor technology, the industry is moving toward more sophisticated, cost-effective, and scalable solutions for autonomous driving

Computing: Demand for computing will increase to process the data and enable motion planning.

Software: The autonomous driving software stack is a layered architecture designed to enable self-driving vehicles to perceive their environment, make decisions, and execute actions. Each of these components is powered by advanced algorithms, machine learning, and real-time computation to achieve reliable and safe autonomous driving. We expect more partnerships and market consolidation as players are building strong position in AD Software Stack.

ODD and HMI are Critical for Safe Adoption and Customer Acceptance

In addition to regulatory requirements, engineering requirements for functional safety will affect the design and operation of modern HMI for automated driving to deliver enhanced safety and take advantage of the side tasks unlocked by higher autonomy.

Consumer education

Understanding what these systems do and their limitations is crucial to avoid misuse and abuse.

Level 3 autonomous driving is not just a technological milestone—it’s a paradigm shift that redefines our relationship with mobility.

Innovation Hubs

Amendment of international regulations as well as national traffic laws will soon give the green light for deployment but the regulatory and legal framework differs across leading car markets. Adoption in emerging markets like China and India is crucial for democratization of Level 3.

References

1. Audi A8 Level 3 System, Audi Media Center ↩︎
2. Honda Traffic Jam Pilot Certification, Honda Global Newsroom ↩︎
3. Mercedes-Benz Drive Pilot System, Mercedes-Benz Official Website ↩︎
4. Auto2x, Advanced Driver Assistance System: Market Growth and Opportunities ↩︎