Autonomy is PinnacleQM's AI and machine learning platform for autonomous vehicle testing in real time. In-vehicle AI executes scenarios from a threat library of close to a million driving hours.
30+years of international
IT experience
Autonomous vehicles must handle endless combinations of road layouts, weather, other road users and local road rules. Physical road testing cannot cover enough scenarios for confidence, and some dangerous hazards are rarely met during test drives at all.
Autonomy processes real-time sensor data in the vehicle and executes test scenarios drawn from a threat library of close to a million driving hours, supported by joint research with Edith Cowan University, extending coverage well beyond physical driving.
Autonomy extends vehicle validation to scenarios that would take lifetimes of driving to meet, giving engineers evidence that road testing cannot.
Of driving scenarios for vehicle testing.
In-vehicle AI processes live sensor data, so perception and responses are assessed as they happen.
Scenarios come from a library of close to a million driving hours, including hazards rarely met on test drives.
Methods are developed through joint university research, so they keep improving with current academic work.
Behaviour is tested against road rules that differ between states, so each market can be covered.
Scenarios repeat exactly, so vehicle software versions can be compared fairly in regression and changes clearly identified.
Recorded results support safety and validation evidence, so engineering and safety reviews are better informed and faster.
Capabilities for validating connected and autonomous vehicle behaviour.
In-vehicle AI processes real-time sensor data during tests, assessing how the vehicle perceives and responds to its environment, so engineers can see exactly where perception falls short and prioritise improvements.
Test scenarios are drawn from a threat library of close to a million driving hours, deliberately including rare but dangerous hazards that physical test drives are unlikely ever to meet.
PinnacleQM works with Edith Cowan University on autonomous vehicle and smart city testing research. Findings from that research flow directly into Autonomy's practical testing methods, keeping them current with academic work.
Results are recorded and organised to support validation and safety evidence, with every finding traceable to the scenario that produced it, so engineering and safety reviews can check exactly what was tested.
Beyond the vehicle, PinnacleQM tests the enterprise, customer and dealer systems that support automotive services, using the same assurance methods as other enterprise programmes, so the whole service chain is assured.
Autonomy engagements are shaped around the vehicle features, scenarios and jurisdictions you need to validate, and the evidence your engineering and safety reviews require.
Every scenario.Agree vehicle features, jurisdictions and validation objectives to shape the scenario plan.
Choose scenarios from the threat library to match objectives and agreed coverage targets.
Run scenarios with real-time sensor processing in the vehicle, capturing data for analysis.
Assess responses against expected behaviour and road rules, and prioritise any issues.
Deliver validation evidence and recommendations, with results retained so future software versions can be compared fairly.
When a core banking platform was upgraded, Enginuity's machine learning updated the automation in 2 hours and every script ran without test pack changes.
Read more
A resources company runs SAP and SuccessFactors regression every month with no testers at the keyboard and $0 additional cost.
Read more
HealthTest obfuscated production records, created full test condition variations and tested more than 5 million transactions per day, without exposing any patient information along the way.
Read more
7 min read
A practical way to decide which regression tests to automate, which to keep manual, and how to measure the return.
7 min read
Many unreliable automated tests fail because of test data and environments, not scripts. How to diagnose and fix the real cause.
7 min read
Tools do not make releases faster by themselves. The practices, ownership and feedback loops that actually change cadence.
Tell us about the vehicle features, scenarios and jurisdictions you need to validate. We will show how Autonomy and our research partnership could support your validation programme and evidence.
Mention the features, jurisdictions and validation stage.
