World Premiere: ZF Annotate Uses AI for ADAS and AD Systems

Modern automated driver assistance systems require a multitude of sensors to precisely analyze the vehicle's surroundings and derive safe driving maneuvers. To advance the development of these advanced ADAS and AD systems, the technology company has developed the cloud-based and AI-supported validation service ZF Annotate. At today's ZF Global Technology Day 2024 in Jeversen near Hanover, the company is presenting the fast, affordable, and market-ready service for cars and commercial vehicles to the media. 

Vehicle needs "Ground Truth"

For the development of advanced and autonomous assistance systems in vehicles, precise and reliable data are essential. Cameras, radar, lidar, or ultrasonic sensors continuously provide information from which the vehicle creates a three-dimensional image of its surroundings. The systems have to recognize various objects in real-time – including, for example, vehicles, people, lanes, or traffic signs. 

These sensor data must be digitally processed correctly so that the vehicle always receives the "absolute truth" – referred to as "Ground Truth" in the professional world – to calculate and implement a driving function based on it. Comparing the collected sensor information with a reliable and highly accurate reference sensor set increases accuracy. This is where ZF Annotate comes in.

AI for Increased Speed and Enhanced Safety

Based on proprietary vehicle data and additional ZF sensor data recordings – the reference measurement – the cloud-based service solution provides the “Ground Truth.” In this process, ZF Annotate acts as an independent, redundant setup separate from the sensor set to be checked, confronted with the same information while driving on the road. The recorded data is then uploaded to the cloud and analyzed. Thanks to artificial intelligence, all relevant objects are accurately marked, classified, attributed, and assigned unique ID numbers. Moving objects are tracked. These object details are part of the complete description of the environmental model, the Ground Truth. After this “annotation,” the software delivers a highly precise comparison measurement. Consequently, ZF Annotate is an AI-powered validation solution for testing and training modern ADAS/AD systems from Level 2+ to Level 5.

“ZF Annotate combines the benefits of a robust and independent reference sensor set with a scalable cloud service that utilizes intelligent 2D and 3D tracking algorithms,” says ZF CEO Dr. Holger Klein.

Previous comparison systems predominantly rely on 2D annotation for the validation of reference data, representing the environment in terms of distance and horizontal angle. The 3D-capable ZF Annotate complements this information by adding the dimension of height.

“This makes our solution a uniquely intelligent ‘cloud factory’ in the industry, optimizing a previously labor-intensive, cost-intensive service in multiple ways,” explains Dr. Holger Klein.

Up to ten times faster and 80% cheaper 

The precise and reliable data from the reference measurements of ZF Annotate can significantly accelerate the development and fine-tuning of complex ADAS and AD systems. Until now, validating such systems was associated with a high workload and was time-consuming and costly because the reference data was traditionally manually annotated by humans. 

“Thanks to Artificial Intelligence, we can accelerate the validation process by up to ten times, reducing annotation time from twelve to two months. This saves our customers time and money,” says Dr. Holger Klein. 

“With ZF Annotate, we are able to generate 'Ground Truth' in the shortest time,” explains Klaus Hofmockel, Head of Research and Development in the Driver Assistance Systems and Electronics division. “With the ability to work 24 hours a day, seven days a week, our cloud-based service completes the validation of reference data in a remarkably short time compared to the market, without any loss of quality.”