Experimental Evaluation of Technology Enablers for Cutting Edge Wearables’ Applications Krystof Zeman 1 , Pavel Masek 1,2 , Jiri Hosek 1,2 , Pavel Dvorak 3 , Radovan Josth 3 , and Tomas Jankech 3 1 Brno University of Technology, Department of Telecommunications, Czech Republic, Brno, Technicka 3082/12 2 RUDN University, 6 Miklukho-Maklaya st, Moscow, 117198, Russia 3 Konica Minolta, Czech Republic, Brno, Zarosicka 4395/13 Contact author’s e-mail: krystof.zeman@phd.feec.vutbr.cz Abstract—Nowdays, wearables are very important part of Internet of Things (IoT) concept. They share a lot of parameters with embedded devices and add some new requirements. Most of those requirements are created by their unique usage - these devices are in user proximity, mostly directly on a human body. This brings also a lot of complications for researchers and engineers. They must consider i.e. safety terms or small dimensions, which are interfering with requirements such as battery life and computational power. This paper presents a study of computational power versus power efficiency of modern System on a Chip (SoC) platform for Internet of Things (IoT). Target of this study is to determine which currently accessible SoC on the market is most suitable for usage in wearables and emerging applications like e.g. augmented reality. Three different platforms were tested and the results are presented at the end of this paper. I. I NTRODUCTION Internet of Things (IoT) is defined as a group of intercon- nected devices. Those devices are communicating with each other without (or with minimal) human interference. Such communication is called Machine-to-Machine (M2M) com- munication. This was a very big change from widely known Human-to-Human (H2H) communication, which is based on interchanging information between people [1]. Thanks to this development, new devices started to appear in almost every field. Most of these devices are small, energy con- strained devices used for measurements and gathering infor- mation [2], [3]. In this paper we will take a closer look at the youngest and most innovative group yet - the wearables [4]. Wearables are defined as devices in near proximity of human body. There are three main directions of current wearables development: smart bands, smart watches and head mounted displays (Aimed to be enablers for the Augmented Reality and Virtual Reality applications). The main reason why wearables are emerging so fast is the wide range of usage. With smart bands it is possible to monitor human health and movement, with smart watches users can get notifications and position information and with AR and VR glasses people can work and learn faster [5]. All of these subcategories of wearables are be- coming more and more resource demanding. In latest forecast, Cisco states there will be 601 million of globally connected wearable devices, producing 335 PB of data traffic [6]. Thanks to this amount of new devices, Juniper Research states that global retail revenue from smart wearable devices will reach 53.2 billion USD by 2019 [7]. As we mentioned before, wearables have very specific requirements. Due to its placement, the form factor is usually very small but it also needs a lot computational power to deliver desired user experience. Currently the most important requirements are very low power consumption, high through- put and computational power [8]. In Table I, there is a comparison of main requirements for each group: The Augmented and Virtual reality devices take a specific role in wearables domain. Augmented reality can be defined as combination of virtual information with real information in a user device (glasses) [9]. This approach needs a lot of computational power, good cameras (one for scene scanning and one for depth scanning) and very high throughput network. For example to stream uncompressed Full HD video (frame rate 60Hz, resolution 1920*1080 with each pixel having three color components), 3Gbps is needed [10]. Most of today available glasses solve these issues by connecting the glasses to the more powerful machine such as computer (META [11]) or computational unit (ATHEER [12]). That means that user is capable of using high definition display units and cameras and to process data going from and to the glasses without losing the user experience. The biggest disadvantage is the cable needed to interconnect those two devices. There are some glasses using only the battery, but they last only about 2-4 hours (Google glass [13], Recon [14]). Currently available AR devices are mostly targeted at industry, education and entertainment purposes. There are a lot of applications for virtualization of objects, guiding through buildings or mainte- nance steps. Those applications are mostly provided by the producer of devices, but it is very common to be able to install third party applications on them. In the future it will be possible to use the AR for almost anything - ranging from everyday usage such as grocery shopping and GPS (Global Positioning System) navigation to 3D development or research. Second very interesting group is Virtual Reality. This tech- nology can be defined as projecting virtual information on a physical display. It means user is no longer attached to real objects and everything he / she sees is virtual. Similarly as AR, the VR needs a lot of computational power. Biggest differences lie in the screen resolution and placement. For VR, the resolution and computational power should be higher than 2016 8th International Congress on Ultra Modern Telecommunications and Control Systems and Workshops (ICUMT) 978-1-4673-8817-7/16/$31.00 ©2016 IEEE 15