Ibeo Automotive Systems, a Hamburg-based automotive LiDAR specialist, has developed a new automation solution for validating ADAS and the sensors needed for autonomous driving.
The ibeo.Reference toolchain also automates and specifies the manual data labeling process. This has traditionally been a very human and time consuming task. You can refer to almost all traditional sensor systems.
ADAS and AD must be validated in a complex process before they are finally mounted on the vehicle and ready for continuous production. Regardless of the technology of the sensor system (camera, radar, or LiDAR), reference data is required for each system and should be compared with the data acquired during the testing process.
Ibeo’s recording system is installed in the vehicle along with the manufacturer’s prototype system to record data in parallel. It can also be used to create scenarios for ADAS or AD algorithm development.
The recorded data collected on the verification drive can be in the thousands to medium hours. In addition, a manual evaluation will generate up to 300 hours of labeling activity for each hour of test drive.
“Manual adjustment of this data is currently a time-consuming and costly factor in many new developments,” said Philip Neitzel, Product Manager, ibeo.Reference.
The industry has traditionally used outsourced service providers to categorize the large amounts of recorded data. Often, hundreds of employees manually evaluate every meter of the road to see if the sensor under test detects a car, person, lane marking, or guardrail, for example to adjust the data. To judge.
Ibeo’s new toolchain, the ibeo Validation Bundle, is working on this task. “As part of the ibeo Validation Bundle, we provide an automated overall solution in the form of ibeo Auto Annotation Software that algorithmically generates reference data from incoming data,” Neitzel added. ..
“If desired, we can manually process these to generate ground truth data, a data that is a realistic and complete representation of the vehicle environment. In this way, one of the major challenges we are facing today. To the series products. “
With the new ibeo.Reference toolchain, users receive the building blocks needed for sensor validation from a single source.
- A recording system consisting of hardware and software to run on reference data. This is the first time that selected third-party LiDAR sensors are also supported.
- As a scalable, cloud-compatible software for automatically creating labeled reference data, ibeo Auto Annotation is an editor for further processing reference data into ground truth data.
“It was very important for us to cover almost every customer challenge in how to combine LiDAR components and to provide a complete system that is technically scaled regularly,” Neitzel said. Stated.
“In this way, we help our customers develop faster, more accurately and in a more cost-effective way, and we are much closer to the overall vision of fully automated mobility.
The ibeo.Reference toolchain is now available. In addition, Ibeo offers a test license, which you can request at support @ ibeo-as.com.
View press release
Original source: https: //www.ibeo-as.com/en
Ibeo is developing ADAS sensor verification | Automotive Industry News
Source link Ibeo is developing ADAS sensor verification | Automotive Industry News