Scenario Mining Tool

Versatile Data Categorization Software for Scenario-Based Testing

Automatically extract, inspect and evaluate concrete driving scenarios existing in your large amounts of recorded or simulated data

Save Time &
Valuable Resources

Find scenarios in large recordings. Avoid wasting time and resources analyzing empty miles. Re-use data.

Acquire Insights &
Evaluate Scenarios

Get scenario KPIs. Apply specific pass/fail criteria. Replay and easily analyse scenarios in detail.

Build Your Own
Scenario Database

Build and grow your own scenario database. Easily integrate 3rd party scenarios from other sources.

Boost Collaboration &
Get Interactive Reports

Collaborate at all company levels. Build and share interactive reports. Proof your covered scenarios.

Focus on What Matters

Find the Right
Scenario Quickly

ADAS/AD development generates large amounts of driving data. However, most of it contains “empty miles” with no relevant driving situations.

IVEX scenario mining software helps you quickly extract relevant driving scenarios buried in your data, to evaluate the real-world driving behavior and potential safety issues of your ADAS/AD functionality.

Seeing is Believing

Inspect any Scenario
with just One-Click

After detecting relevant scenarios, easily inspect your scenario of interest using a single click. Visualize, replay and share the scenario with other colleagues for faster ADAS/AD failure detection.

Easily evaluate the performance of your ADAS/AD system through the analysis of relevant scenarios based on a number of metrics including the time-to-collision and time-to-headway.

Simple Integration

Easy Toolchain Integration Enabling end-to-end Data Processing at Scale

IVEX scenario mining tool can ingest real-world or simulated data in any format. This versatility makes it easy to integrate into your existing autonomous vehicle development toolchain, either on-cloud or on-premises.

The tool can automatically process new arriving data and it enables re-processing of existing datasets at wish. IVEX software handles all the data orchestration, processing and escalation.

Traceability

A Traceable Analytic Solution

Every processing step the data takes is traceable, and therefore replicable. IVEX scenario mining tool helps to create a safety argumentation to use during your test, validation and homologation processes.

For example you can provide scenario coverage reports supported by easily accessible data, to early stop your collection campaigns when requirements are met, or to prove there were no gaps in your testing campaigns.

Novelty

Latest State of the Art Methodology

The tool gives you a structured methodology derived from the latest research on Scenario-Based Testing and applicable regulations such as EU 2019/2144, EU 2022/1426, standards such as SOTIF, ISO 26262, ISO 34502, as well as internal requirements.

Scenario Mining Steps

Step 1 – Input driving data

IVEX software can receive either already perceived scenes (e.g. Bounding boxes and track IDs); or raw driving data (e.g. LiDAR, localization) to which we apply our IVEX raw to 3D perception algorithms. Customizable data processing pipelines are also available, for example including blurry plates or faces modules, ODD extraction, etc.

Step 2 – Logical scenario search:

The scenario mining tool scans the input driving or simulation data and extracts pre-defined logical scenarios (e.g. Cut-In, Approaching Stationary Vehicle, etc). These scenarios come from regulations such as EU 2019/2144, EU 2022/1426, standards such as SOTIF, ISO 26262, ISO 34502, etc) as well as internal requirements.

Step 3 – Extract relevant parameters:

For each scenario detected, the tool extracts its specific parameters (e.g. Initial Velocity, Target Position, etc) to obtain the concrete scenario faced. At the same time, scenario KPIs (e.g. Time to Collision, Time Headway, etc), valuable metadata and ODD (e.g. road type, curvature radius, weather, etc) can be extracted from the specific segment of data. Customized KPIs can also be added at this stage (e.g. Time to Safe Distance).

Step 4 – Concrete scenario evaluation:

For example, a design requirement may specify that the reaction of the vehicle to a Cut-In scenario should provide a safe distance to the new leading vehicle in less than 3 seconds. The user can define a criteria indicating that any scenario where “Time to Safe Distance greater than 3 seconds” is a failure.

Step 5 – Scenario Inspection & Coverage

Access each interesting scenario and inspect its parameters and KPIs evolution. Identify issues and gaps within your scenario coverage. From here several options are available: share specific scenarios with colleagues; build rich and interactive reports; enlarge your scenario database; download the segment specific of raw data; close the loop by re-simulating specific scenarios, etc.