Back
x min
14.05.2025

Releasing autonomous software faster with DeepScenario and TNO's StreetWise

For the safe deployment of autonomous driving, a deep understanding of real-world traffic is required. This means that detailed knowledge of all traffic situations in an operational design domain (ODD) must be gained. To achieve this, large volumes of data need to be collected and analyzed. However, traditional data acquisition with prototype vehicles is not only costly, but also extremely time-consuming. On top of that, advanced software tools are needed to analyze the recorded data.

To address these challenges, DeepScenario and TNO, the Dutch research organization Netherlands Organization for Applied Scientific Research, work together to provide an industry-leading, data-driven solution for the safe deployment of autonomous driving. This solution is already in use by Torc Robotics for autonomous trucking [1], [2] and has also been adopted by other companies in the automotive sector.

Unlocking global access to high-quality traffic data with DeepScenario

DeepScenario’s AI software equips customers with mission-critical capabilities to virtualize videos from any monocular camera, including dashcams, traffic cameras, or drones. Users can upload large-scale data campaigns to DeepScenario’s web platform, where they are processed by the company’s proprietary computer vision software. Their 3D virtualization pipeline detects and tracks all dynamic objects in the scene with centimeter precision, providing crucial information like positions, dimensions, and velocities for data analysis. The processed data can then be accessed and downloaded directly from the web platform. By supporting both moving and stationary camera setups, this versatile technology not only unlocks the full potential of fleet data, but also massively accelerates data collection at critical locations like intersections. Remarkably, the stationary approach is 100x faster compared to dedicated measurement vehicles for targeted scenario acquisition.

Getting maximum insight from trajectory data with TNO’s StreetWise

StreetWise is a powerful dataset characterization tool that helps users understand which traffic situations are represented by trajectory data in these datasets. This is a crucial step in ODD modeling for requirement setting and safety validation. Moreover, it supports data curation and coverage visualization for machine learning components in the autonomous vehicle. Leveraging real-world data to identify relevant scenarios for specification, testing, training, and safety validation, and foregoing the issues of “boring miles” in traffic data or simulating unrealistic scenarios saves enormous amounts of time and cost and therefore provides maximum return on investment on data acquisition.

Understanding the driving environment in detail with StreetWise

Introducing a holistic approach for the deployment of autonomous driving

The collaboration between DeepScenario and TNO opens up new possibilities for the cost-effective and efficient development of autonomous software. Through this synergy, trajectory data extracted by DeepScenario can be directly fed into TNO’s StreetWise to identify scenarios for data driven specification, development, and validation. The combination of both solutions leads to faster development cycles and accelerates the release of autonomous systems.

Interested users can test this powerful toolchain by downloading already processed traffic data from DeepScenario’s web platform and requesting a demo of TNO’s StreetWise.

We use cookies to enhance your browsing experience, evaluate website usage, and serve marketing content. Visit our Privacy Policy for more information.