Guest Blog: How IoT Will Transform Wearables

In this blog from member company ElikoAlar Kuusik, R&D Manager and Senior Researcher at the Tallinn University of Technology, offers his insights on the growth of IP-based connectivity for the internet of things (IoT), how it can be utilized for IoT wearables and the need for smart object interoperability.

IP-based connectivity is the communication mechanism most commonly used for modern IoT applications – such as Industry 4.0, and smart city and smart agriculture sectors.

IP networking is also gaining attention in the sectors that, traditionally, have relied on non-IP connectivity, such as smart homes and in-vehicle communication. The communication standards for smart homes and home automation were developed more than 15 years ago. At that time, IP-stack capable microcontrollers were more expensive and more rare than devices that supported lighter networking protocols such as ZWave or Zigbee. For in-vehicle communication, low-level CAN-bus communication was sufficient because the external real-time connectivity was not a consideration. Now seamless connectivity and end-to-end security is crucial for both smart homes and smart vehicles, which is pushing these verticals rapidly towards IP.

One of the last ‘smart sectors’ still not largely based on IP is wearables. Previously, wide area network connectivity for wearables was not considered particularly important. Today the situation is completely different: The majority of wearable sensors are Internet connected and use smartphones as gateways. Even modern heart pacemakers have wireless connectivity options for device configuring. The communication between a smartphone and back-end server is naturally IP-based. End-to-end IP connectivity would further simplify the sensor data exchange and improve security.

Bluetooth Low Energy (BLE) is easily the dominant technology for low-power short-range wireless communication today. Unfortunately, with some proprietary exceptions, BLE did not support IP communication and multi-hop networking until last year. This also meant the IP networking of wearables wasn’t possible. However, the new BLE version 5.0 will be standardized in 2017. Besides faster communication and longer distance it will support IP-based addressing, enabling IP-compliant networking of wearables.

Wearables will rely on Fog instead of Cloud

Fog computing is an interesting new data handling strategy. Instead of today’s central processing of data in clouds, fog computing relies on local and distributed data processing. Fog computing improves the system reliability and reduces the load on communication channels. However, the local data processing resources, or amount of available information, are usually not sufficient to avoid cloud services completely. Health and wellness applications essentially require connectivity solutions  that simultaneously support local ad-hoc connectivity for fog-like data processing as well as cloud interfacing. The local ad-hoc connectivity is necessary for reliable and real-time decision-making, which, in relation to health and wellness applications, can literally mean the difference between life and death.

As stated above, the cloud interfacing element simultaneously provides connectivity – with personal health record repositories, for instance. In the future, patient medical devices need to communicate with conventional hospital medical equipment or wellness training equipment in the gym – and preferably directly, rather than via a centralised service. Moreover, emerging applications, such as eyesight-related wearables, may require the option to ‘communicate with’ traffic signs or public transport vehicles. Low-power BLE communication is perfectly suitable for prompt ad-hoc connectivity – it is present in every smartphone, iBeacon and Eddystone localization device. IP networking, on the other hand, enables smooth cloud connectivity and seamless data exchange with “thick” computers and other Internet enabled devices that have any kind of physical interface. It is highly likely that IP-addressing combined with BLE 5.0 low-power radio connectivity will be the industry standard for the next generation of wearable devices.

IP-based communication enables better semantic interoperability

Besides connectivity, data interoperability issues have to be addressed as well in the context of wearable devices. The health and wellness sector has long battled semantic interoperability issues. Although there is a range of well-developed medical taxonomies and terminologies, such as SNOMED CT and ICD, they are weak in realizing machine-to-machine interoperability or composing personal health records. For example, with byte-level predefined Bluetooth LE GATT structure it is possible to present systolic and diastolic blood pressure values but not to differentiate between ventricular pressure measured inside the heart and ventricular pressure measured using a cuff.

It is highly unlikely that new and universal predefined data descriptions and relevant configuration options for the health domain can be developed because we don’t yet know what kind of information needs to be transmitted in the context of emerging personal medicine or lab-on-a-chip. In future, instead of being restricted by GATT health and wellness profile fields, IP networking protocols such as CoAP will enable exchange of detailed – but still flexibly structured – information. The latest medical data exchange standard FHIR (by Health Level Seven International) supports an efficient and freely extendable JSON data format. FHIR messages can be composed with SNOMED CT 9-digit object classifiers, achieving an extremely high level of data precision with fractional overheads and maintain the interoperability with existing health information systems.

IoT communication is bound to conquer the last standing non-IP “fort” in the wearables domain. As well as achieving greater flexibility and semantic interoperability, IP-communication – in combination with a new BLE 5.0 wireless communication standard – will improve the reliability and security of health and wellness data.