Short answer: a single-point 905 nm rangefinder isn’t a LiDAR replacement — it’s the right tool for the specific jobs LiDAR is overkill for: forward collision distance on a fixed axis, fork-height sensing, dock and charging-pad alignment, drop-off detection. Keep 2D or 3D LiDAR for 360° navigation and mapping, and let a single-point module handle the point measurements a full scan doesn’t need to make.
Where AGVs and AMRs actually need this
(If you’re solving the equivalent problem in the air rather than on the ground, see our 905 nm vs 1535 nm guide for UAV gimbals and payloads — the wavelength trade-offs carry over, but the mission profile doesn’t.)
Three scenarios come up constantly in warehouse and yard automation, and none of them need a point cloud:
- Forward collision distance. An AMR moving down a picking aisle needs one number, updated fast: how far to whatever’s directly ahead. A bumper- or chassis-mounted single-point sensor answers that continuously without processing a scan.
- Fork-height and pallet-face sensing. A forklift-style AGV lifting to a specific rack level needs a fixed-axis distance reading to confirm fork height or pallet-face position — a job that doesn’t benefit from a 360° field of view.
- Dock and charging-pad alignment. Docking against a charging station or a trailer edge is a one-axis alignment problem: distance-to-target along the approach vector, checked at high update rate as the vehicle closes in.
Single-point rangefinder vs 2D/3D LiDAR: the real trade-offs
Cost and complexity. A mechanically scanning LiDAR needs a rotating or MEMS beam-steering mechanism and more powerful emission/detection electronics to return usable signal across a full scan — bill-of-materials cost and moving parts a fixed single-point module doesn’t carry. Time-of-flight sensors are “significantly more affordable” than other 3D depth-scanning technologies and can undercut most LiDAR units on unit cost, per an industry overview of ToF sensors in mobile robots (The Robot Report) — that source gives no exact figures, so treat it as a qualitative signal, not a number to quote a customer.
Update rate. A commonly cited 2D LiDAR scan rate for AMR navigation is 25–40 Hz, full 360° rotation — Hokuyo’s UST-10LX publishes 10 m range, 0.25° resolution, 40 Hz (Hokuyo USA). A single-point module measures one fixed axis repeatedly rather than scanning a field, so it’s not directly comparable — but it is faster: LRF22VB is specified up to 200 Hz, LRF50VB up to 100 Hz (sources: LRF22VB.md, LRF50VB.md), a quicker reaction loop than waiting on a full LiDAR rotation back to the same bearing.
Range and coverage. LiDAR wins on coverage by definition — 3D LiDAR can generate “tens of thousands or even millions of points per second” (Hokuyo). A single-point module only reports distance along the axis it’s pointed at: LiDAR answers “what’s around me,” a single-point sensor answers “how far is the thing I’ve already decided to look at.”
Reflectivity robustness. This is a physics constraint both technologies share, not a LiDAR-specific advantage or a rangefinder-specific weakness — and our own datasheets make the effect explicit. LRF50VB is rated 0.05–50 m against a 90%-reflectivity target but only 0.05–20 m against a 10%-reflectivity target; LRF22VB drops from 0.05–20 m to 0.05–10 m across the same reflectivity range (sources: LRF50VB.md, LRF22VB.md). Plan around your worst-case obstacle — a dark rubber wheel stop or a matte-black pallet edge — not the datasheet’s best-case number.
Where the two complement each other. Hokuyo’s own guidance ties 2D LiDAR to indoor warehouse navigation meeting the ISO 3691-4 driverless-industrial-truck standard, and 3D LiDAR to loading docks, yards and overhead-hazard environments — neither is what a single-point 905 nm module is for. The architecture most AMR integrators land on is LiDAR for navigation and mapping, plus single-point modules at the specific fixed points on the chassis — bumper, fork, dock sensor — where a fast-update, low-power, no-moving-parts measurement beats waiting on the next scan cycle.
Selection parameters that actually matter
- Update rate (Hz). Match to how fast the platform closes on an obstacle — a slow AGV in a warehouse aisle doesn’t need 200 Hz, but a faster-moving AMR’s collision sensor benefits from it.
- Minimum range. This family starts at 0.05 m, which is what makes it usable for close-quarters jobs like fork-height or drop-off sensing that a long-reach, 1 m+-minimum module can’t do (source:
LRF22VB.md,LRF50VB.md). - White/black target range delta. Size your requirement against the low-reflectivity datasheet number, not the high-reflectivity one — see above.
- Interface. LRF22VB and LRF50VB both support UART and I²C (source:
LRF22VB.md,LRF50VB.md), so pick whichever bus your controller has spare capacity on. - Supply voltage. LRF22VB: 3.0–3.6 V DC. LRF50VB: 3.3–5 V DC. LRF100H, at 100 m range, steps up to 9–36 V DC, suited to a larger AGV’s power distribution rather than a board-level 3.3 V rail (sources:
LRF22VB.md,LRF50VB.md,LRF100H.md).
Integration notes
Mounting angle. Mount square to the surface or object you’re measuring, and account for the beam’s field of view (roughly 1.9–3° across this family) when sizing minimum standoff.
Sensor fusion, not sensor replacement. A single-point rangefinder is one clean input into a stack that typically already includes encoders, an IMU, bump switches or a camera — it contributes the one thing none of those give cleanly: a calibrated distance along a known axis. Our short-range DToF for industrial & robotics article goes deeper on this framing.
Protocol and drivers. All modules output UART-TTL, readable with a standard serial driver on an STM32, an Arduino-class board, or a Raspberry Pi — no proprietary SDK. Frame format and a minimal parse loop are in our UART-TTL integration guide. We’re also publishing open-source drivers for common robotics platforms, including ROS2 — see the erdi-lrf-drivers repo.
Recommended models
| Model | Range (90% / 10% reflectivity) | Update rate | Interface | Supply | Weight | Fits |
|---|---|---|---|---|---|---|
| LRF22VB | 0.05–20 m / 0.05–10 m | up to 200 Hz (selectable) | UART / I²C | 3.0–3.6 V DC | 1.5 g | Short-corridor forward collision, compact AGV mounts |
| LRF50VB | 0.05–50 m / 0.05–20 m | up to 100 Hz | UART / I²C | 3.3–5 V DC | 1.3 g | Longer aisle/open-warehouse sensing, dock alignment |
| LRF100H | 0.05–100 m / 0.05–30 m | 20 Hz–10 kHz (default 1 kHz) | UART / I²C | 9–36 V DC | 20 ± 2 g | Larger AGVs on a higher-voltage rail, longer-range yard sensing |
Sources: LRF22VB.md, LRF50VB.md, LRF100H.md.
Frequently Asked Questions
Can a single-point 905 nm rangefinder replace 2D LiDAR on an AMR?
Not for 360-degree navigation and mapping — that’s a job 2D LiDAR is built for and a single fixed-axis beam isn’t. For a specific job like forward collision distance, fork-height sensing, or dock alignment, a single-point module is usually the better-fitted tool: lower cost, lower power, and no moving parts, at the price of only reporting distance along one axis instead of a scan.
What update rate does a 905 nm single-point module give compared to LiDAR?
Our own short-range modules run well past typical 2D LiDAR scan rates on a single axis: LRF22VB outputs up to 200 Hz and LRF50VB up to 100 Hz, against a 2D LiDAR’s commonly cited 25–40 Hz full-rotation scan rate. That’s not an apples-to-apples comparison — LiDAR is scanning a full field of view each cycle, the single-point module is only ever looking at one axis — but for a fixed collision-distance check, the single-point module updates faster.
Does target reflectivity affect AGV rangefinder range the way it affects LiDAR?
Yes, for both technologies — it’s a DToF physics constraint, not something specific to either sensor category. Our own datasheets show it directly: LRF50VB is rated 50 m against a 90% reflectivity target but only 20 m against a 10% reflectivity target, and LRF22VB drops from 20 m to 10 m across the same reflectivity range. Dark, low-reflectivity obstacles always cut effective range versus a bright target — plan your worst-case obstacle, not the datasheet’s best case.
What interface does a 905 nm rangefinder module use to connect to a robot controller?
UART-TTL across our catalog, readable by any microcontroller or single-board computer with a standard serial driver — an STM32, an Arduino-class board, or a Raspberry Pi, with no proprietary SDK required. See our UART-TTL integration guide for the frame format and a minimal parse loop.
Are open-source drivers available for ROS or Arduino integration?
We’re publishing open-source drivers for common platforms, including ROS2 — see the erdi-lrf-drivers repo.
Get a datasheet or a sample quote
Tell us your AMR’s obstacle-avoidance job — forward collision, fork-height, dock alignment, or something else — along with your worst-case target reflectivity and available power rail, and we’ll confirm the right model and current lead time in writing.