Simulate, Validate, Repeat: How Apex.Grace Streamlines Testing with Record, Replay & Simulation
- Apex.AI
- Jun 18
- 2 min read
Updated: Jun 30
As autonomous systems mature, the ability to test, validate, and debug them efficiently has become just as critical as the algorithms themselves. In real-world applications, engineers face a hard challenge: ensuring safe, predictable behavior across an endless number of possible scenarios, while only testing a subset.
That’s why Apex.Grace was built to make simulation, record & replay, and time control native to the development and validation process. Whether you’re running early-stage algorithms or evaluating production-ready software, Apex.Grace provides a seamless path from simulation to real-world deployment without rewriting application logic.
A Unified Testing Pipeline Across MiL, SiL, and HiL
Apex.Grace supports a flexible testing and integration pipeline that spans:
Model-in-the-Loop (MiL)
Software-in-the-Loop (SiL)
Hardware-in-the-Loop (HiL)
What ties it all together? The ability to feed applications with live data, recordings, or simulation inputs interchangeably. That means the same software you test in simulation can run unchanged on a vehicle, greatly accelerating iteration and de-risking deployment.
And with Apex.Grace’s virtual time support systems can execute faster or slower than real time, while preserving correct behavior and deterministic timing.
Record Once, Replay Forever
Built on ROS 2’s rosbag2 framework—and actively shaped by Apex.AI’s contributions—Apex.Grace takes recording and replay to the next level:
Selective data capture from sensors or messages exchanged between applications
Distributed recording across multiple ECUs with merge and sync support
Plugin-based architecture for efficient serialization (CDR), compression, and backends like SQLite and MCAP
This enables highly scalable recording strategies that capture real-world data in structured formats, ideal for regression testing, debugging, and reproducing field issues.
Deterministic Replay for Reproducible Results
In complex systems, even subtle timing variations can introduce noise into testing and debugging. Apex.Grace solves this with Deterministic Replay—a core capability that ensures consistent, reproducible behavior during playback.
It supports:
Fixed-order Replay: Forces a strict execution sequence, regardless of how the system originally ran. Perfect for regression and integration tests.
Reproduced-order Replay (coming soon): Replays the original runtime behavior exactly, ideal for post-mortem analysis or incident reproduction.
These replay modes are coordinated through Replay Executors and a centralized Replay Coordinator, which precisely control timing and task order across the system.
Simulation-Ready, Out of the Box
Whether you're validating early concepts or refining corner-case behavior, Apex.Grace integrates smoothly with leading simulation platforms:
CARLA and LGSVL (ROS 2-compatible, open source)
Proprietary, production-grade simulators used by OEMs and Tier 1 suppliers
A Simulator Bridge ensures seamless translation between Apex.Grace’s DDS-based messages and simulator-specific protocols. This enables:
Realistic sensor data input
Full control loop testing with feedback into the simulation
Accurate synchronization via virtual time
Simulated scenarios can now feed directly into your system as if they were live, supporting rigorous validation under unsafe, rare, or complex conditions—before you ever turn a wheel on a test track.
Why It Matters
Testing is no longer a stage. It’s a continuous loop of simulation, validation, and improvement.
With Record, Replay & Simulation, Apex.Grace brings deterministic control, scalable recording, and simulator integration into one cohesive development workflow. It removes barriers between test environments and production, empowering engineers to iterate faster, test smarter, and deploy with confidence.
Whether you’re debugging a localization edge case or validating behavior at scale—Apex.Grace helps you simulate reality before you face it.
