by Neil Banerjee , Chief Evangelist, AVCC
The automotive industry has undergone significant changes with the advent of autonomous and electric vehicle (EV) technologies. Despite advances in artificial intelligence (AI), sensor technology, and compute capabilities, the journey towards full autonomy (SAE Level 5) is fraught with technical, regulatory, and market challenges. This paper explores the technological evolution of the automotive ecosystem, the complexities inherent in higher levels of vehicle autonomy, and the strategies employed by original equipment manufacturers (OEMs) to navigate this landscape. The role of consortia like the Autonomous Vehicle Computing Consortium (AVCC) in addressing these challenges through collaboration, standardization, and benchmarking will also be examined. 
Introduction
The global automotive landscape has seen a dramatic shift over the past two decades, with electric vehicles (EVs), autonomous driving systems, and software-defined vehicles (SDVs) leading the way. However, the adoption of these technologies has been slower than expected, particularly for fully autonomous vehicles (SAE Level 5). The 2012 breakthrough in deep learning (DL) for perception systems fueled optimism, but significant barriers—ranging from technical limitations to regulatory hurdles—have impeded widespread adoption.
We aim to provide a comprehensive analysis of the challenges faced by automotive OEMs and suppliers in launching new vehicle technologies, with a focus on autonomy, software integration, and system complexity. Additionally, we will highlight the contributions of the AVCC in mitigating these challenges through industry-wide collaboration.
Technological Landscape and Innovations in Vehicle Development
Evolution of Vehicle Technology
The automotive industry has traditionally been hardware-driven, with mechanical and electrical components forming the core of vehicle systems. Over the last 25 years, a shift towards digital convergence has emerged, integrating advanced software, AI, machine learning (ML), and sensor technologies into the vehicle’s architecture. Key technological developments include:
- 1997-2016: The foundation of modern electric vehicles (EVs) and autonomous driving systems, with semiconductors, connectivity, and embedded systems playing an integral role.
- 2016-2024: The rise of AI/ML in perception systems, with cloud computing, industrialization, and digital twins supporting vehicle manufacturing and lifecycle management.
Despite advancements, autonomous vehicles (AVs) remain in the early stages of consumer adoption. While sensors and compute capabilities have significantly evolved, these have not yet led to mass adoption of fully autonomous vehicles (Level 4/5).
Complexity in Vehicle Autonomy: The SAE Levels 1-5
The SAE International defines autonomy across five levels, with each step increasing the complexity of the vehicle’s architecture and reliance on software and sensor fusion for control.
Overview of SAE Levels
| SAE Level | Function | Description |
| Level 1 | Driver Assistance | Basic driver assistance features like adaptive cruise control (ACC) using sensors like cameras. |
| Level 2 | Partial Automation | Multiple automated functions operating simultaneously; human supervision required. |
| Level 3 | Conditional Automation | Vehicle controls most functions; human intervention needed only in specific conditions. |
| Level 4 | High Automation | Autonomous operation in designated environments; full control over all driving tasks. |
| Level 5 | Full Automation | Complete vehicle autonomy in all environments without human intervention. |
Detailed Breakdown of Automation Levels: Costs and Technology
| SAE Level | Sensors | Compute Power | Software | Redundancy Systems | Cost Range |
| Level 1 | Basic sensors (cameras/radar) | Minimal | Basic algorithms | None | $200 – $2,000 |
| Level 2 | Mid-range radar/multiple cameras | Advanced processors | Complex algorithms | Some redundancy | $2,300 – $5,800 |
| Level 3 | High-res cameras/radar/LiDAR | AI chips | AI-based algorithms | Multiple fail-safes | $6,500 – $13,500 |
| Level 4 | Advanced sensor array | High-performance AI computing | Advanced AI algorithms | Full redundancy | $12,500 – $32,000 |
| Level 5 | Sophisticated sensor array | Extremely high-performance AI | Deep learning algorithms | Full redundancy | $30,000 – $80,0 |
Challenges in System Integration
Software-Defined Vehicles (SDV)
In 2024, Software-Defined Vehicles (SDVs) have emerged as a critical aspect of vehicle development, enabling OEMs to push updates and introduce new features without physical modifications. However, SDVs also introduce new complexities, particularly in cybersecurity and data management.
Homologation and Certification
Homologation remains a key barrier to vehicle deployment with stringent certification requirements across regions (e.g., U.S., EU, China). Autonomous vehicle systems must meet specific performance and safety criteria to ensure regulatory compliance.
AVCC’s Role in Standardizing and Accelerating Autonomy
Mission and Objectives
The Autonomous Vehicle Computing Consortium (AVCC) focuses on reducing complexity by promoting standardization and collaboration:
- Specifying system architectures for AV computing
- Defining interconnects for AV system components
- Collaborating with OEMs and suppliers for interoperability
Value Creation through Collaboration
As no single OEM or supplier can independently tackle the immense challenges of developing fully autonomous vehicles, AVCC fosters industry-wide collaboration by standardizing non-differentiating technologies.
Conclusion
The journey towards fully autonomous vehicles remains one of the most significant technological challenges. While advancements have revolutionized the automotive landscape, much work remains to achieve full autonomy. Consortia like AVCC play a pivotal role by fostering collaboration to overcome these challenges.
References
- McKinsey Report – “Rethinking Car Software and Electronics Architecture” – Feb 2018
- Deloitte US and Cars.com Automotive Reports (2024)
- Car Edge Analysis (2024)
- AVCC.org Membership Information (2024)
