NEON Location Service Architecture

Sarah Blaszak -

NEON sensor fusion, ranging, and mapping algorithms deliver location and user context inside buildings, in subterranean structures, and in other GPS-denied areas even where ranging, inertial, magnetic and pressure sensors can provide erroneous measurements due to interference from the environment and sensor drift.

NEON Software conforms with evolving commercial platform architectures which include three layers: sensor hub processors, application processors, and server or cloud-based processing:

  • Sensor Hub Software performs sensor processing using higher sample rate data. Functions performed include user context determination, user motion modeling, step detection, step length/step direction computation, confidence estimation, feature detection, and calibration. The software runs on a dedicated accessory (e.g., the NEON Tracking Accessory) or on a cell phone platform using embedded sensors.
  • Navigation Engine Software produces an absolute location (X, Y, Z) and provides location error estimation. The technology enables easy and accurate conversion of multi-rate and non-Gaussian measurements (e.g., GPS, BLE, landmark/feature matching, external sensors) into constraints on the localization solution. Functions performed include drift detection and removal, compass calibration, constraint filtering (e.g., GPS, BLE ranging, external constraints), indoor outdoor detection, elevation processing, and application of landmark/feature constraints. The Navigation Engine runs on the Android application processor.
  • NEON Location Assistance (NLA) Server holds map information including a priori data such as satellite base maps, elevation maps, terrain elevation data, beacon locations, waypoint locations, and building models, as well as inferred map information such as reference pressure and structural, magnetic, and RF features. Map tools are available to support both manual input and bulk import of known map data. An automated map fusion tool processes and fuses crowdsourced map features and signatures required to develop 3-D navigation maps. The Location Assistance data is stored in an Azure cloud, and location assistance processing is performed within the cloud.

Location Services Architecture

Discovered map information is sent to the NLA server where it fuses the crowdsourced map features and provides a central repository for that data as well as for pressure reference, terrain and other map information. A user connects to the NLA server to load available map data but once the map data is downloaded it will persist and further connection to the NLA server will not be necessary. When any mapped features are encountered, those map constraints are processed by the Navigation Engine Software improving global and relative location.

 

 

Have more questions? Submit a request

Comments

Powered by Zendesk