Page 42

EETE APR 2014

WEARABLE electronics Eco-system management for a smart sensing textiles By Régis Latawiec The wearable electronics market is already valued in billions of dollars and is promising to boom over the next decade. Driven by the healthcare sector, wearable devices also target fitness and wellness applications. Systems typically track steps and body position using accelerometers (MEMs), heart rate beats using pulse sensors, etc. Today these devices are usually connected to a smartphone or to smartwatches through wireless communication channels such as Bluetooth. The move towards a more global usage of wearable technologies requires the sensors to be directly embedded into textile and connected to a wearable gateway through a Body Area Network (BAN) using various communication means including human body as a transmission medium. This pervasive strategy of globally connected clothes will enable new usages and create new business opportunities in connection with body care monitoring, healthcare monitoring (mobile health, aka m-health) and also extreme conditions activities such as professional sport activities, especially in harsh environment like high altitude and high seas. The typical architecture for wearable electronic systems implies four main blocks: Sensors, a gateway, a smart device, and the cloud – see figure 1. Sensors capture real-time physiological parameters from various parts of the body. As they need to be small and extremely low power the sensor end-points have very low computation capabilities. The data is collected by a gateway and processed locally (data fusion) to provide instant information to the user on a smartwatch for instance. When connected to a smartphone, the collected BAN data is also sent for complex data monitoring or sent to the Cloud for future data management Fig. 1: The four blocks of a typical architecture for a wearable electronic system. Fig. 2: A wearable BAN with various services attached to it. and post-analysis. There are many challenges involved in that architecture. One of the most obvious addressed today by textile and semiconductor manufacturers is to integrate sensors and electronics to fabrics in a way that clothes remain fashionable and easy to manufacture. The next challenge is about providing application services to the end user. This means to embed more intelligence in clothes, to build an open and secure eco-system to collect, analyze and process data, and to provide end users with augmented information. Let’s explore the wearable gateway architecture to provide such intelligence and eco-system infrastructures. A data gateway Like sensors, the gateway needs to be small and lightweight (less than 30g), highly integrated and power efficient. For these reasons, microcontrollers (MCUs) are the best choice for designing the electronic systems, compared to application processors (MPUs). When considering software architectures, the wearable gateway needs to be customizable, just like a smartphone is. It has to offer the capability to execute new software application services designed by different manufacturers from the eco-system and to rule the way these different applications can cooperate between each other (data sharing, access 3rd party sensors, etc.). For instance, one buys a smart T-shirt including its gateway and connected sensors. When buying other clothes with new sensors from the same vendor or from another vendor, the gateway should allow the download of new software applications related to those new sensors – see figure 2. To implement the above use-case, gateways must provide capabilities like supervisor/user execution modes, dynamic linking and resources monitoring to deal with hardware resource access right management or access right between services (software component) from different 3rd parties of the eco-system. They must also provide a clean API and a SDK including a simulator for the software applications developers. Like most application platform infrastructures (Android and Windows for instance) , the way to design such an execution platform relies on a technology that provides virtualization means and a clean, easy to use standardized component framework. The proposed architecture relies on a full Java embedded Virtual Machine tightly coupled to a supervisor that monitors two execution modes: Kernel mode and Feature mode. This Java environment offers embedded libraries, including custom ones dedicated to the gateway’s eco-system. It supports life-cycle management of the Features, which are organized as software components (OSGi™ is usually used as a component framework). Finally, the “Application Domain Kernel” provides Régis Latawiec is COO at IS2T - www.is2t.com – He can be reached at regis.latawiec@is2t.com 42 Electronic Engineering Times Europe April 2014 www.electronics-eetimes.com


EETE APR 2014
To see the actual publication please follow the link above