“The complexity of this technology and scenarios in which it must operate vastly exceeds anything previously attempted in this industry.”
- Forbes, August 1, 2018
Measurable
Safety
Coverage Driven Verification
Foretify includes:
(1) A Measurable Scenario Description Language (M-SDL) used to describe both scenarios and coverage goals at a very high level (soon to be made open) to enable a ‘measurable safety’ ecosystem
(2) Automation to generate combinations of combinations of scenario variants along with monitors to check and track scenario coverage
(3) The ability to aggregate coverage data from the execution across all testing platforms (test vehicles, test tracks, simulation, X-in-the-loop) and extract analytics and metrics for developers to know where they are, what to do next and when they’re done
'One to Many'
Condensed
‘One’ Behavioral Specification of Scenario Categories May Represent 100s of Thousands of Scenario Variations
Readable
Specifications are Human Readable and Can be Easily ‘Audited, Adapted, and Augmented’
Adaptable
Only Scenario Specification that Can be Easily Extended and Morphed for New Operational Design Domains (ODDs)
'Many to One'
Aggregation
A Hierarchy of Monitors Aggregates Coverage Data from 100s of Millions of Scenario Variants
Analytics
Coverage Holes are Identified and Highest Impact Tests are Ranked Using Causal Analysis
Metrics
% Coverage of Any Level in Scenario Category Hierarchy is Measured and Reported via Annotation of the vPlan
The Coverage Driven Verification Flow
Powered by Foretify
- Define a verification and coverage plan for all scenario categories (for example complex lane changes, sensors faults, weather conditions and more)
- For each scenario category use an open, high level scenario description language to define scenario behaviors and their measurable coverage requirements including the mixing of scenario categories essential to reach the challenging edge cases
- Automatically generate 100’s of millions of meaningful core + edge cases scenarios, coverage monitors and graders
- Analyze the results
- Measure coverage of each scenario category in the verification plan
- Identify coverage holes in each scenario category
- Rank tests according to the contribution to coverage
- Evaluate the grading of scenarios according to different metrics (comfort, safety buffers, rules etc.)
- Much more…
- Iterate, adapt, refine and improve the coverage results to reach the desired safety confidence level
