Indoor Navigation
Infrastructure Inaccessibility
After nearly half a century of developments in satellite navigation, we can now achieve centimeter accuracies where we have a clear view of the sky. That’s no help indoors, however, where GNSS signals cannot be received. Many different indoor mapping methods have been proposed and tested, including emitted RF ranging (using WiFi, RFID, pseudolites, Bluetooth beacons, or other beacons of opportunity), SLAM (simultaneous localization and mapping), LiDAR, close-range photogrammetry, pulsed signals from lights, magnetic distortions, and sonic imaging. They require some form of infrastructure (such as WiFi hotspots) that must be mapped prior to use and are often not available in buildings under construction or dilapidated buildings, or installing and revisiting targets, which is one of the pain points of static surveying.
An Out-of-the-Box Solution to Indoor Mapping
Applanix chose to tackle this challenge by building on its decades-long expertise in inertial navigation. However, an inertial system inherently suffers from drift. How can you constrain it without pairing it with a GNSS receiver or first surveying an indoor space by traditional methods?
Applanix’s answer, the Trimble Indoor Mobile Mapping Solution (TIMMS), combines a mechanical approach with a software one. To aid its inertial navigation solution, its rigid wheels prevent it from moving sideways and are instrumented to provide a very precise measurement of its forward motion. It then uses a point-cloud data adjustment (PCDA) software — a SLAM-type algorithm — to analyze the point cloud and find commonalities in different passes. As a result, the inertial system only requires a single precise starting point. For larger buildings, quick traverse though the building is recommended to establish multiple points with high relative accuracy. The TIMMS’ current spec is about 2 cm, but under controlled circumstances it has achieved sub-centimeter accuracy.
Fully-Equipped, with No Stitching
The TIMMS is superior to traditional methods in another key way: it outputs a single, completely registered, colorized point cloud for the entire space surveyed, which eliminates the need to stitch together hundreds of discrete scans. It also outputs georeferenced spherical images. Because the camera is rigidly mounted to the cart, the system knows exactly where it was and where it was pointed every time it took a photo.
Each TIMMS cart holds a 3D laser scanner, a FLIR Ladybug panoramic camera, 2 wheel encoders, an inertial measurement unit (IMU), as well as data storage and an operator display. Before each scanning run, the team initializes the system using a control point and then pushes the cart in a systematic way through the building.
To view the data, Applanix recommend Trimble MX Software. This suite of tools provides a full solution for remote access of the TIMMS data such that a user can view the panoramic images and interact with the underlying point cloud to make measurements and/or identify and map assets for GIS or inventory management applications. Applanix is also working with Trimble RealWorks, to ensure all traditional LiDAR capabilities and workflows are supported.
Quick Results, even in Congested Environments
Using a TIMMS, an Applanix team captured comprehensive 3D scans and spherical imagery of 1.75 million sq. ft. of floor space in two terminals at Los Angeles International Airport in just 32 hours — compared to up to six weeks of field work and about as long for processing and modeling that it would have taken for a conventional static scanning survey. After scanning the terminals, the Applanix team used the TIMMS Post Processing Suite to create 3D point clouds and panoramic images. The final deliverable, which had a positioning accuracy of 1 to 2 cm, included high-density point clouds, clouds at 1cm and 5cm density, GIS data, 2D floor plans, and 3D models.
The Latest in Post-Processing
Recently, Applanix released the TIMMS Spatial Processor 2.0, an upgraded companion post-processing software which offers: a completely redesigned user interface, support for a greater number of CPUs for faster processing, and new GCP event editors. With this software, users can now: globally change coordinates, datums and timing of GCP observations; select from an exhaustive list, or define custom map projections; and save processed data separately from real-time data, among other benefits. View the TSP 2.0 Infosheet for a breakdown of the software's key components.
Sources
Kevin Andrews, Director of Indoor Products, Applanix, interviewed by Matteo Luccio on 2017.06.01
John Stenmark, “Indoor Mobile Mapping Takes Off at LAX,” The American Surveyor, 2017 January