Pallet Docking
Documentation Home / Pallet Docking
SmartDocking Pallet Docking
Pallet entry-face position, orientation, and status for autonomous forklift docking.
Before you update: Check your installed AW3 and algorithm versions against the software history below. PalletPro users should use the legacy downloads and guide for their deployment.
Designed for forklift pallet positioning SmartDocking locates the pallet fork-entry face using 2D semantic positioning or pure 3D geometry. It supplies a reference position, orientation, and positioning status to the vehicle. AW3 manages devices, configuration, calibration, and result monitoring; the vehicle controller selects targets and controls motion and safety.

Release status
Latest documentation
V0.1 · 2026-09-28
User Manual · White Paper
Latest algorithm package
SmartDocking 3.0.2
2026-09-29 · RKU20 / Linux ARM64
SmartDocking Package · Release notes
Required frontend
AW3 Installer
Frontend only; no AW3 PC backend
View software history ↑
View document history ↑
How it works
2D semantic positioning combines RGB model recognition with depth data. Pure 3D positioning matches leg and beam geometry against a configured template. Select and calibrate the required mode; pure 3D is an alternative method, not an automatic fallback when 2D fails.

Outputs and capabilities
Entry-face reference position
X, Y, and Z in millimeters for the pallet fork-entry reference point.
Orientation
Roll and Yaw in degrees; Pitch is fixed at zero and is not a valid measured output.
Positioning status
Check validity, detected count, and available pallet-class results before vehicle use.
Two positioning modes
Select an appropriate semantic model or pure 3D geometry template for the site pallets.
Pallet templates and calibration
Maintain pallet geometry and installation-specific settings.
Load detection
Available in pure 3D mode when enabled; it is not a general 2D output.
Validate the installed camera, pallet types, entry-face visibility, calibration, and complete vehicle docking behavior. Model coverage and geometry settings must match the site's pallets; this page does not claim universal recognition or collision safety.
Installation
RKU20, Linux ARM64
Deploy the .tar.gz to the matching device / center via SSH, following the User Manual.
Both packages are required. Do not install the AW3 PC backend for this application. The algorithm archive contains the device-side service and runtime dependencies; it is not a Windows installer. Confirm the device platform before updating and back up its configuration. Hardware deployment has not been tested as part of this publication.
Start here
1. Review application fit and operating limits
2. Install the AW3 frontend and the device algorithm package
3. Configure, calibrate, and validate the application
Last updated