This box suggests the current best answer
Prior revisions of answers
Proposal 1 (discussion item )
Migrating from TMC to OLR: what are typical issues:
CoPilot input
Great question! Migrating from TMC (Traffic Message Channel) location referencing—which uses pre-coded locations—to an on-the-fly method like OpenLR introduces several technical and operational challenges. Here are the key issues:
First present fundamental reasons to migrate:
Benefits
Opposite argument: TMC coverage is limited to pre-coded locations, OpenLR is unlimited
1. Fundamental Differences in Referencing Approach
- TMC: Uses a fixed, pre-defined location table (with IDs for road segments, junctions, etc.). This ensures consistency but limits flexibility.
- OpenLR: Encodes locations dynamically based on geometry and attributes, making it map-agnostic and more scalable.
Implication: Migration requires rethinking how locations are identified and transmitted—no more reliance on static IDs.
2. Map Provider Variability
- Different providers (Google, HERE, TomTom) have different network topologies, segment definitions, and attribute sets.
- OpenLR is designed to be map-independent, but accuracy depends on how well the encoding matches the target map.
Challenge: Ensuring interoperability and consistent decoding across providers.
3. Accuracy and Ambiguity
- TMC guarantees exact matches because of pre-coded IDs.
- OpenLR uses shortest-path algorithms and attribute matching, which can lead to:
- Ambiguity in dense urban areas.
- Errors if road attributes differ between source and target maps.
Mitigation: Fine-tuning encoding parameters and implementing robust fallback strategies.
Opposite argument: TMC coverage is limited to pre-coded locations, OpenLR is unlimited
4. Performance and Bandwidth
- TMC: Compact messages (just IDs).
- OpenLR: Larger payloads (geometry, attributes), which can impact bandwidth and processing time.
Consideration: Efficient encoding and compression are critical for real-time systems.
Size: Less of an issue now going from ~2-4 bytes to ~20-30 bytes per location reference
Encoding / decoding time and processing impact:
- Encoding is more compute-intensive, but most often done on large server
- Decoding on user devices is much more compute intensive, and especially in large urban areas with lots of traffic messages (e.g. London, Paris) may strain a CPU and take a long time to process, which is a risk if the next batch of traffic messages.
5. Legacy System Integration
- Many traffic systems, navigation devices, and broadcast protocols are built around TMC.
- Migration requires:
- Dual support during transition.
- Conversion tools (TMC ↔ OpenLR) for backward compatibility.
6. Quality Assurance
- Need to validate that OpenLR-referenced locations match intended real-world positions in a map-agnostic way ( handling multiple map versions from same map vendors, as well as maps from different vendors).
- Requires thorough testing and needs SW and data infrastructure for automatic machine-based error detection.
- Variations in maps and map updates could impact accuracy, and re-validation is advice.
7. Governance and Standardization
- TMC is standardized and widely adopted, but supporting various versions of TMC location tables presents large overhead at the server side/encoding side.
- OpenLR is open source, also widely adopted, but less regulated, so encoder/decoder compatibility may require attention.
Comments
| 1 | |||
| 2 | |||
| 3 |