Internet of Things (IoT) sensors can collect huge amounts of data. One of the key challenges for a design engineer is deciding how to manage all this data. Even though it's possible to collect immense amounts of data, this approach is not always ideal because of unnecessary data management costs that would be incurred. A more economical approach to IoT sensor data management is to employ local, embedded data analysis to filter very large data samples and send alerts only on a significant change in the data. This strategy was employed in an IoT prototype, developed by Honeywell and Intel Corporations, utilizing pattern-matching technology.
Called The Honeywell Connected Worker Solution, this proof-of-concept IoT prototype is a personal connected safety solution for factory laborers, mine-workers, first responders, firefighters and other industrial personnel. The solution monitors workers for toxic gas exposure, breathing, heart rate, posture, and motion. The resulting data and actionable intelligence is displayed remotely on a visual, cloud-based dashboard, giving plant managers and incident commanders the information needed to better anticipate unsafe conditions and prevent potential “man-down” scenarios that could threaten worker safety.
The Connected Worker Solution consists of Honeywell’s Self-Contained Breathing Apparatus (SCBA), which sends sensor data via short-range wireless, Bluetooth Low-Energy (BLE) to a wearable mobile hub for sensor fusion and data transmission via Wi-Fi or cellular to the cloud for data ingestion via a Trusted Analytics Platform and visualization on an AWS-hosted central incident command application. The wearable sensors are based on the Intel Quark SE microcontroller, a low-power edge processor, which provides a sensor hub combined with pattern-matching technology for sensor data stream processing and pattern recognition.
The pattern matching technology feature available on the Intel Quark SE microcontroller is unique in that it's not programmed using traditional instruction code; rather, it is loaded with a memory array that embodies a set of sensor-derived feature vectors that allow a highly parallel set of processing elements to classify live sensor data streams against familiar patterns.
The pattern matching technology plays an important role in both gesture-based communication and activity identification for the Connected Worker Solution. When initially programmed, the Intel Quark SE microcontroller gesture device is trained with a dictionary of gestures (e.g., the letters A, D, M, L, and U) for detecting and reporting back in real-time. For activity identification, the Smart Body Activity Device is similarly loaded with a dictionary of motions and activities that would provide meaningful insight for remote reporting and visualization. When such a letter is recognized, the pattern matching technology alerts the processor in the gesture device that the specific letter has been seen. The host processor then communicates the gesture event detected over BLE to the mobile hub.
By handling the low-level sensor processing and gesture recognition locally on the Intel Quark SE microcontroller-based wearable device, wireless communication channel traffic can be reduced (from device-to-wearable and from mobile-hub-to-cloud) by avoiding the need to transmit a streaming feed of raw sensor data to the cloud for remote processing and gesture analysis.
This allows for real-time and localized feedback in an efficient manner and reserves wireless bandwidth consumption for sensor fusion applications. The pattern matching technology also provides another advantage: it offloads its own host processor from having to perform gesture and activity processing in software. This allows for powerful pattern analysis while ensuring the necessary long battery life expected from edge sensing devices and additional memory and computing resources for the resident host application.