How reliable data can enable safety in ADAS

To advance AVL’s Dynamic Ground Truth reference system, AVL needed two key things:

  • Quick Sensor Calibration
  • High accuracy labelling, delivered fast

The Challenge

  • Traditional sensor calibration used by AVL has many inherent issues.
  • Existing Data annotation tools and services providers were not good.

Deepen Solution

  • High quality data collection through calibrated sensors.
  • Deepen’s calibration suite cuts the time spent on multi-sensor calibration from hours to seconds.
  • Deepen’s annotation technology brings AVL more efficient, precise, and accurate annotations.

The Result

  • With Deepen Calibrate, AVL was able to accurately calibrate sensors in seconds.
  • The DGT recorded ground truth data offering an extremely accurate view of the vehicle and its systems’ test environments.
  • AVL was able to precisely annotate large number of frames. The data included 12 object classes with multiple attributes.

Executive summary

Advanced Driver Assistance Systems (ADAS) is revolutionary technology that could change the nature of transportation by increasing car and driving safety. The scope of ADAS is huge:

ADAS incorporates dozens of technologies, including:

  • Lane Keeping Assist
  • Braking Assist
  • Blind Spot Indication
  • Adaptive Cruise Control
  • Parking Assist
About a third of all new cars sold in major markets have ADAS features, and that number is only growing

ADAS relies on the fact that the system can accurately perceive its environment and respond to it. In order to do its job, ADAS relies on data on the vehicle’s surroundings, and lots of it. This data is constantly streaming in through a handful of sensors on the vehicle, such as numerous cameras, radars, sonars, and perhaps most importantly, light detection and ranging (LiDARs).

This data needs to be of high quality for the algorithms to be accurate to ensure safety.

AVL Challenge

Perception of the environment is both a crucial and a challenging task for Advanced Driver Assistant Systems (ADAS) and Autonomous Driving (AD). A wide range of sensors, including LiDAR, cameras and radars are used to build the perception layer of these systems and ensure a precise view of the surroundings in diverse traffic and environmental situations.

In order to enhance data accuracy for ADAS & AD systems, AVL has developed Dynamic Ground Truth:

  • High resolution measuring system used to capture reliable environment data
  • Box mounted on vehicle w/ lidar, camera, and high precision GPS sensors
  • Lets you capture 360 degree field of view.

To advance AVL’s Dynamic Ground Truth reference system, AVL needed two key things:

  • Quick Sensor Calibration
  • High accuracy labelling, delivered fast

AVL’s problem statement

Data integrity is at core of AVL’s DGT system. AVL’s problems were two fold:

Traditional sensor calibration has many inherent issues:

  • Time consuming
  • Expensive
  • Not scalable

Existing Data annotation tools and services providers were not good enough:

  • Inexperienced labeling team
  • Lack of custom engineering
  • Poor data operations and lifecycle management

Deepen AI ensures data integrity

Step 1 - Deepen Calibrate - High quality data collection through calibrated sensors

Deepen accomplishes this through its web-based calibration suite

  • If sensors aren’t calibrated → then all the data they collect is less accurate → everything the data are used for (e.g. teaching the machine) becomes less reliable
  • Deepen calibrates a variety of sensors (LiDAR, radar, camera, etc) to quickly provide a complex, detailed, and accurate reading of the data used.

Deepen’s calibration suite cuts the time spent on multi-sensor calibration from hours to seconds.

Step 2 - Deepen Annotation - Accurate data annotation

Deepen’s annotation technology brings AVL more efficient, precise, and accurate annotations. One of the most important uses of this data is to train ML models autonomous systems can learn from to learn how to navigate the world. In order to do so, we must annotate, fuse and label the data we have collected.

Annotation types supported:

  • Bounding boxes
  • Semantic segmentation
  • Polylines
  • Scenario labeling
  • Keypoints
Download to Read Full Sase Study
CALIBRATION

Deepen Calibrate manages the complexities, ensuring accuracy and making autonomous systems safer.

Accuracy

Same or better accuracy than traditional calibration

  • Sub 1-degree angular error
  • Centimeter-level transfer error
  • Pixel-level re-projection error

Customizable & Scalable

Our web-based sensorcalibration solution allows for up-to-date easy scaling across all key calibration types

  • Customizable for all sensors
  • Quick and simple set-up
  • No need of updating expensive hardware every year

Cost Efficiency

Brings down cost by more than 60-70% as compared to traditional sensor calibration

  • Intrinsic calibration
  • Extrinsic calibration
  • Supporting - LiDAR, Camera, Vehicle, IMU, Radar & more

Annotation - The Deepen Advantage

Deepen is the only safety first annotation solution that allows you to label every 2d pixel & 3d point for thousand of frames at scale. Supporting all state of the art labeling features

Pixel and Point Level Handling

Only ADAS and AV focused tooling company that handles pixel & point level for thousands of frames.

Custom Engineering Team

Dedicated and exclusive frontend, backend, AI, Product and Project management teams.

Flexible workforce models

Advanced Reporting

To support all business needs to give clients key information with a click of a button

Lifecycle task-allocation and Management

Complete lifecycle of tasks made to support huge annotation teams, with the ability to streamline and automate your workflow experience

Quality Assurance

Built-in QA Flow allows enterprises to easily verify the quality of processed data to maintain quality standards with features like sub-sampling, comments, and automatic checker

Automation

Ai Sense is our smart labeling function that pre-labels the data to increase efficiency & reduce man-hours by more than 85%

“Independent, high-precision ground truth reference systemssuch as AVL DGT play an increasingly role for our OEMs toobjectively develop and validate their vehicles sensor performance.

With Deepen Calibrate, we can ensure that our AVLDGT-sensors are calibrated with high accuracy in a fractionof time during driving campaigns”

Thomas Guntschnig,
Product Portfolio Manager ADAS/AD AVL

Result

AVL’s Dynamic Ground Truth Reference System (DGT) leveraged the sensor calibration suite Deepen Calibrate, to enable precise sensor calibration during driving campaigns and accurate ground truth data collection. AVL also used Deepen AI’s annotation tools and services to get the data ready to be used for safe development and validation.

Calibration

With Deepen Calibrate, AVL was able to accurately calibrate sensors in seconds. The DGT recorded ground truth data consisting of lidar, camera, and high-precision GPS sensors, offering an extremely accurate view of the vehicle and its systems’ test environments. This enabled engineers to make statistical analysis of large amounts of representative, accurate real-world driving data to assess the performance of the ADAS/AD sensors and their perception algorithms.

  • 90% fewer man hours
  • 85% lower cost

Annotation

With Deepen annotation tools and services, AVL was able to precisely annotate large number of frames. The data included 12 object classes like bicycle, motorcycle, truck, trailer, car, person, etc with multiple attributes.

In addition, Deepen AI ensured GDPR compliance.

  • GDPR compliance
  • 40% time saving
  • Custom feature development

Conclusion

Reliable data is the foundation for safety. By combining our efforts at Deepen AI through Deepen Calibrate & annotate and AVL’s cutting-edge AVL Dynamic Ground Truth™ solution, we can significantly increase the quality of data and increase the safety of ADAS/AD systems.

Meet us at CES

from 5th to 7th Jan 2022
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.