The Role of Low Fidelity Simulations in ADS Validation
Before moving to sensor-rich, high-cost simulation environments, validating the core decision-making logic of an Automated Driving System (ADS) in a structured, lightweight manner is key. Low Fidelity simulation offers this capability by operating on object-level data, focusing on how the system plans and responds without the noise of raw sensor data.
This stage is critical for ensuring that the planning algorithms, driving policies, and control modules are performing correctly before moving on to more intensive stages of testing.
Running Low Fidelity Test Suites in ReplicaR
With Low Fidelity Execution, ReplicaR allows engineers to run any test suite—built from MyScenarios and AutoScenarios—through object-level simulation. The ReplicaR execution engine ensures robust, consistent, and efficient runtime, while focusing solely on the behavior of dynamic objects in the scenario.
Key features include:
- Granular Simulation: The system-under-test (SuT) interacts with an object list rather than simulated sensors, enabling direct inspection of how it processes world states and generates actions.
- Scenario Reuse: All previously curated scenarios in TestSuites can be directly loaded and simulated, enabling regression and functional testing with minimal setup.
Planning and Policy Validation
At this stage of validation, the main goal is to analyze how well the system:
- Interprets object inputs
- Plans paths
- Executes control logic
This is where planning and behavior policies are tuned for reliability, safety, and predictability. Any undesired behaviors can be caught early—before entering high-fidelity or real-world testing.
Data Handling and Result Access
Simulation inputs and outputs are processed using the ASAM OSI format, supporting standardization across the toolchain. Results are stored in ReplicaR’s secure database for traceability and further analysis via the ScenAnalytics module.
Conclusion
Low Fidelity Execution plays a foundational role in the simulation stack by enabling fast, scalable, and focused validation of ADS planning behavior. It reduces cost, increases iteration speed, and ensures your ADS software is fundamentally sound before diving into complex simulation layers.