Real Time Location Detection and Monitoring System (RTLS)

Solution overview

Our end-to-end solution allows real-time registering and processing of position information of goods (stock item), heavy machinery (assembly line), vehicles (forklift) or people (doctor, patient). Our architecture enables real time data visualization, as well as batch data processing with machine learning algorithms.

Positioning
We use the unique SUNSTONE real-time localization system based on UWB (ultra-wide band) radio that enables to track devices indoor with 10-50 cm (4-20 inch) precision. Other radiofrequency localization technologies, such as Wi-Fi and Bluetooth work with significantly lower precision, which does not allow exact position detection required in many cases, for example identifying goods in a large warehouse or in a factory.

The SUNSTONE RTLS is provided by OMTLab (Hungary) Ltd. The infrastructure consists of three main elements. Tag is a small (3x3x2cm) battery-powered element, which needs to be attached to the item to be tracked. Anchor is a wall unit performing the measurements. At least 4 anchors needed for one zone. Central Unit manages the Anchors, calculates positions and makes the data available online.

Features: 

• Precision ~10 cm
• Processing in 0,05 seconds
• Range 100m ~ 2000m2 per zone, hundreds of tags

Data Processing
Data processing and storing are performed on a big data architecture. Our architecture consists of two data processing layers:

Speed Layer is able to capture all sensor data in real time. A Kafka cluster ensures scalability and secures that no data can be lost, while the necessary transformations and data enrichment is done with StreamSets. Processed information is stored in Solr collections, which is efficient in case of real time dashboarding. Speed layer collects all available information, but it is designed to process only those, which are really needed to be displayed in real time. The rest is transferred to the batch layer.

Batch Layer is able to receive data from traditional sources via scheduled batch jobs or to stream data from the speed layer for storing data for long term. The information is stored in Impala tables, which is transformed, aggregated and enriched here. Batch layer is also the place for advanced analytics and machine learning algorithms.

Due to Cloudera features, storage capability and processing performance can be elastically scalable and works with cloud, on-premise or hybrid environments.

United Consult’s solution is able to provide browser-based, real time dashboards with Banana’s open-source technology representing the actual position of hundreds of tags. The pre-aggregated and structured data can be further analysed with Qlik or any other modern, self-service BI tool. Event-based raw data or pre-processed data can be exported for various reasons, for example establishing alerting functionality.

Features: 

• Real-time data processing
• Real-time data visualization
• Elastic scalability
• Standard processes and data model