From Vector Maps to Simulation-Ready: Creating 3D Environments from Compact HD Maps and Exporting Across Formats

Simulation environments need structured, semantically rich maps that support scenario generation, localization, and motion planning under controlled and repeatable conditions. HD maps provide this foundation: Their object-based road models encode geometry, connectivity, and regulatory semantics with centimeter-level accuracy, providing a simulation backbone for scalable, high-fidelity testing. They allow engineers to generate edge cases around occlusions, crossings, and nondrivable regions while also supporting ego-pose correction in degraded GNSS zones and feeding route logic into behavioral planners. Because HD maps are schema-driven and versionable, they integrate seamlessly across toolchains—from SIL/HIL simulation to deployment—reducing engineering friction and enabling continuous feedback loops with fleet telemetry.

While HD maps are essential, the lidar-based HD map pipeline is resource-intensive and the point-based information can be redundant for many automated driving (AD) and advanced driver assistance system (ADAS) applications, especially when deployed at scale. Leveraging a mid-level representation is essential to create a balance between detail and efficiency. Like full HD maps, compact HD vector maps capture drivable space and roadside elements as structured geometries (e.g., points, lines, and polygons) enriched with semantic metadata. This object-oriented format allows for fast querying, targeted updates, and modular expansion without the overhead of denser lidar-based mapping. Compact HD maps serve the same core AD/ADAS functions—localization through visual landmarks, lane-level motion planning, and simulation via semantic anchors for automatic scenario creation—while being efficient enough to deploy across large urban geographies.

This article highlights how to transform compact HD vector maps into simulation-ready content compatible with RoadRunner and other platforms (Figure 1).

RoadRunner HD Map Format

RoadRunner HD Map (RRHD) is a documented, generic HD map format from MathWorks. The RoadRunner HD format is specifically designed for high-definition map data, focusing on detailed representations of road layouts, lanes, lane boundaries, lane markings, junctions, signs, barriers, and other scene elements. The data is serialized using protocol buffers and saved in a binary format with the .rrhd extension. Protocol Buffers (protobuf), developed by Google, is a language-agnostic binary serialization format that ensures efficient data storage and transfer. RRHD maps can be compiled using languages such as Python^® and C++, or directly manipulated using MATLAB^® objects and functions through the RoadRunner HD Map MATLAB API, which provides programmatic access to create, edit, and validate map elements such as roads, lanes, junctions, and signals. Please refer to the “RoadRunner High-Definition Maps: A Guide to Creating RRHD with MATLAB API” white paper for more details on the API capabilities and integration workflow.

Additionally, RRHD is built for seamless integration with simulation tools such as RoadRunner, MATLAB, and Simulink^®. This makes it straightforward to use in the development and testing of automated driving systems.

RoadRunner supports custom map imports, allowing GeoMate to serialize GeoMate HD maps using protocol buffers and bring them into RoadRunner Scene Builder, where the generated 3D scenes can be exported in ASAM OpenDRIVE^® or Filmbox^® formats for simulation and testing.

RoadRunner is a powerful tool for generating realistic, reusable driving environments and scenarios that significantly accelerate and enhance the development and testing of ADAS and AD technologies.

Converting Vector Maps to RRHD

Multiple proprietary vector maps for HD maps have been created. Compact HD vector maps can be translated into RoadRunner-compatible .rrhd files by aligning each map element with RoadRunner schema. This process maps lane geometries, road boundaries, and roadside objects—such as signs, crosswalks, and poles—into semantically enriched representations suitable for simulation. Complex logic, including traffic signal phases, turn restrictions, and routing rules, should be handled to ensure that traffic behavior in the resulting scenarios is both realistic and reproducible. To scale this conversion, both MATLAB APIs and MATLAB scripting should be leveraged, allowing preprocessing, transformation, and export automation across diverse geographies. This enables the delivery of simulation-ready maps that seamlessly integrate with RoadRunner and other AD development platforms.

Validating the .rrhd File in RoadRunner

The validation of the .rrhd file in RoadRunner is driven by RoadRunner’s built-in automated tools. RoadRunner validates the map by flagging issues such as broken lane connectivity, missing signal links, or invalid road geometries. At GeoMate, the validation process relies on these system-generated errors and warnings to guide any necessary adjustments, minimizing manual debugging. Validation confirms that the file adheres to the expected schema—ensuring correct organization of roads, lanes, intersections, and metadata—and that semantic elements such as turn restrictions, signal phases, and route logic are accurately preserved. This automation-first validation loop efficiently ensures the map’s integrity and readiness for simulation while significantly reducing the time required for manual inspection and tuning (Figure 2).

Exporting to Multiple Simulation Formats

RoadRunner provides interactive editing and sign placement tools, along with native export capabilities to OpenDRIVE (.xodr) or FBX formats for integration with a variety of simulators. A seamless workflow from HD map data to .rrhd supports efficient integration with MATLAB, Simulink, and RoadRunner. This allows engineers to iterate quickly, from ground-truth data ingestion to SIL/HIL testing, while avoiding the storage and processing overhead associated with point cloud–based mapping.

Real-World Applications and Simulation Impact

Organizations like VUFO use RoadRunner to simulate crash scenarios from real accident data for safety research. For example:

• VUFO simulates crash scenarios from real accident data to support safety research.
• Aptiv reconstructs sensor-based driving trajectories to test ADAS algorithms in virtual environments.
• Toyota builds scalable, cross-platform 3D road models to streamline simulation workflows and reduce development time.

RoadRunner supports high-fidelity scene creation with lane-level accuracy, dynamic traffic modeling, and compatibility with industry-standard formats such as ASAM OpenDRIVE and OpenSCENARIO. These capabilities enable faster scenario generation, reduced testing costs, improved validation accuracy, and broader reuse across simulation platforms, accelerating innovation in vehicle safety and autonomy.

Conclusion and Future Work

The presented end-to-end pipeline—from compact HD vector map ingestion to .rrhd generation and validation—streamlines the creation of simulation-ready content for ADAS, autonomous driving, and smart city planning. By combining schema-aware transformation, automated validation via RoadRunner’s tools, and support for multiple industry-standard formats, this workflow significantly reduces the engineering overhead traditionally associated with HD map generation.

This pipeline will further accelerate simulation preparation, improve coverage across diverse urban geographies, and support rapid iteration in AD development environments and safety validation (Figure 3).

For those interested in deeper technical details, we recommend exploring the RoadRunner product page, GeoMate’s RealSimE platform, and the ASAM OpenDRIVE standard for interoperability.