A new lidar system developed by researchers at the University of Toronto, the Vector Institute, Ciena, and POSTECH points toward a practical shift in 3D sensing: richer information from each measurement.
The system, described in the Optica paper “Polarimetric Full-Wavefield Coherent Lidar,” measures depth, velocity, and polarization properties at the same time. In plain terms, it can help determine where an object is, how fast it is moving, and how light interacts with its surface. That last piece opens the door to material-sensitive sensing, which could make lidar more useful in robotics, autonomous vehicles, remote sensing, and industrial inspection.
Richer Data From Each Lidar Return
The practical value comes from reducing ambiguity.
Most lidar systems are excellent at geometry. They tell a machine where surfaces are located and help build a 3D understanding of the world. Other sensors often carry the burden of interpretation. Cameras add visual context. Radar adds motion information. Thermal cameras add heat signatures. Specialized sensors can provide material clues.
That kind of multimodal stack can work well, but it adds complexity. Each sensor must be mounted, powered, calibrated, synchronized, processed, maintained, and fused into a usable perception model. Every added sensor also adds another point of failure and another source of disagreement.
This research suggests another path. Instead of relying entirely on separate sensors to resolve uncertainty, the lidar return itself can carry more information.
That does not mean a single lidar will replace every other sensor. It means lidar may be able to shoulder more of the perception workload.
Simplifying the Sensor Stack
The researchers’ project page describes the system as a “polarimetric full-wavefield coherent lidar” that simultaneously measures depth, velocity, and polarization. The system builds on capabilities found in coherent optical modem technology, which is already used in high-speed telecommunications. Those modems are designed to measure light with extraordinary precision, including amplitude, phase, frequency, and polarization. Repurposing that technology for sensing gives lidar access to information that conventional systems typically leave behind.
For autonomous vehicles, that could help with confidence in difficult scenes. A vehicle does not simply need to know that something is in the road. It needs to understand whether the surface is a sign, a wet patch of pavement, vegetation, plastic, metal, clothing, or another vehicle. The source coverage notes that polarization data revealed sign lettering and helped distinguish artificial vegetation from real vegetation in ways intensity data alone did not.
That kind of added confidence can improve decision-making in edge cases, especially where lighting, glare, weather, or clutter make visual interpretation harder.
For robotics, richer lidar could help machines interact more effectively with the physical world. A robot in a warehouse, farm, factory, hospital, or disaster zone needs to understand objects well enough to avoid, sort, grip, inspect, or manipulate them. Geometry helps with location and shape. Material-sensitive information could help with judgment.
Better Outcomes From Smarter Sensing
Industrial inspection may be one of the clearest near-term opportunities. If a scanning system can collect geometry, motion, and surface-sensitive information together, it could support better inspection of coatings, corrosion, wear, roughness, contamination, or surface change. The strongest business case may come from reducing repeat inspections, catching problems earlier, and turning more field data into actionable maintenance decisions.
Remote sensing could also become more valuable. A 3D dataset that includes surface-response information is more useful than a geometry-only point cloud in applications such as vegetation mapping, infrastructure monitoring, road condition assessment, disaster response, and environmental change detection.
There are still practical limitations. The researchers note that further work is needed on bandwidth, streaming acquisition, and data transfer for faster dynamic scenes. Those are important engineering barriers between a research demonstration and a widely deployable commercial system.
Even so, the direction is significant. The next wave of lidar innovation may come from getting more value out of every return. Better sensors do not always need to mean larger sensor stacks. In some applications, better sensing may mean fewer blind spots per measurement.
For industries trying to deploy autonomy, robotics, inspection, and remote sensing at scale, that is the larger promise: lidar that does more than map shape, and starts helping machines understand physical conditions.
Read More Recent News: Gaussian Splatting and LiDAR: A Practitioner’s Field Guide
