To meet emerging regulations and get lifesaving active-safety features into the hands of more drivers, OEMs need one thing above all else in their sensing and perception systems: flexibility.
Safety regulations around the world — such as the European Union’s General Safety Regulation 2 (GSR2), which will roll out new requirements in 2024 — are mandating certain advanced safety features across vehicles. When features like automatic emergency braking and emergency lane keeping are incorporated into new vehicles, they will go a long way toward reducing accidents and fatalities.
Many of these safety features can be built into vehicles by adding a minimal number of sensors, which means that the cost of vehicles that include them should remain within reach of most consumers. For example, the features mentioned above can be implemented with a vision system consisting of a single camera, either alone or in conjunction with forward-facing radar.
The challenge is that some vision-based perception solutions on the market bundle a wide range of capabilities into their systems without providing the flexibility to optimize the capabilities for required features. That means an OEM could have to pay for capabilities such as free-space detection or high-beam controls even if they are unrelated to the safety features the OEM wants to offer on a particular model. In addition, some vision solutions are tied to a specific system-on-a-chip (SoC), so the OEM has no latitude to decide which provider’s SoC is included in its vehicle.
When OEMs have limited ability to specify the system design or software functions that would meet their requirements, they face challenges scaling safety capabilities across their product lines.
A fresh approach
The answer is to move to an open and flexible platform of software and systems — one that is SoC-agnostic, does not force the OEM into set bundles of software capabilities, and can scale from vision-only safety features related to GSR2 compliance to a Level 2+ or above automated driving solution that fuses 360-degree sensing from multiple sensing modalities.
Aptiv’s Gen 7 Vision solution meets all of those criteria. It fulfills GSR2 requirements while supporting NCAP 2025 up to a 5-star rating. It enables vision implementations that are optimized for specific requirements, resulting in the best overall system cost without sacrificing performance, and allowing OEMs to affordably scale their vision systems from entry-level vehicles through premium models. OEMs also have flexibility when it comes to selecting software capabilities and an SoC for the system.
The forward-facing camera itself is available in two primary versions: a smart 3-megapixel camera that would most often be used for Level 0 to Level 2 automated driving, and an 8-megapixel satellite camera most commonly used for Level 2+ and beyond.
As OEMs look ahead to greater functionality and higher levels of automation, the system can easily integrate vision data with radar data for sensor fusion, providing a comprehensive model of the environment around the vehicle. Radar performs in all weather and lighting conditions, with improved object detection and packaging flexibility, while maintaining low power and compute requirements.Flexibility is key to innovation. To develop highly optimized, high-performance perception systems, OEMs will need solutions that meet the specific needs of their customers today while providing opportunities to scale to more robust features in the future.