An MQTT-based Context-aware Wearable Assessment Platform for Smart Watches
Abstract ConceptualSmart watches offer a one of a kind chance to sanction universal learning situations. The different implicit sensors in these watches give logical data that can be specifically consolidated into pervasive learning structures. This paper introduces the structure and usage of a wearable appraisal stage dependent on the Samsung Gear S2 Smart Watch. Executed utilizing the IoT MQTT convention, this stage enables the framework to consequently buy in to and distribute inquiries to understudies dependent on their present setting.
For instance, understudies will get informed on their savvy with an inquiry identified with their area or circumstance got from the savvy sensor information. The framework utilizes the PONTE MQTT server and is coordinated with Google’s Classroom.
keywords Crowdsourcing, Internet of Things, QTI, MQTT , Sensor Context, Smart watch, Wearable.
The utilization of wearable gadgets has developed exponentially in recent years. Starting at now there are 441 wearable gadgets signed in the vandrico database.
These wearable gadgets have been ended up being reasonable in setting of training. For instance, getting input while going to an address by noting inquiries without intruding on the address through message, or in situ direction by accepting criticism while performing exercises in class. Labus et al.  investigate the utilization of savvy watches to push suggestions to understudies from a Learning Management System (LMS).
A key favorable position of shrewd looks for learning is the capacity to incorporate ‘setting mindfulness’ in learning structure. Position and climate of student are basic instances of setting. We can order setting into essential or auxiliary. Essential setting comprises of perusing the sensor information and removing data straightforwardly from the sensor. Auxiliary setting, then again, is the setting when the data separated is utilized to determine another dimension of data, for example, getting the interesting identifier of an article from its RFID, and inferring the qualities of the item by perusing from a database regarding its ID. For instance, for a savvy, the accelerometer and spinner information is essential, while construing whether the stationary, strolling or running is the auxiliary setting.
We can separate setting life cycle into four stages; setting securing, demonstrating, thinking and spread. Obtaining, thus, can be activated from programming (pull) or the sensor equipment (push), both push and draw can happen either or interims or in a split second by a particular choice. Procurement can be immediate, through setting server, or through some middleware. The sensors utilized amid obtaining can be physical, virtual (e.g., twitter status) or intelligent (e.g., an administration speaking to a huge number of sensors, for example, climate). It ought to be noticed that for a savvy an eventbased securing is increasingly proper because of intensity utilization contemplations. For the demonstrating stage, setting can be displayed utilizing basic key esteem sets, organized XML, social tables, ontologies with explicit area extensions or JSON. The thinking stage comprises of three stages; preprocessing the information, for example, evacuating anomalies, sensor information combination to consolidate readings from numerous sensors and setting derivation by creating data from the information gathered and have experienced the two stages, for example, getting area names from GPS organizes. Thinking can be done utilizing distinctive models like if/else rules, regulated adapting, for example, choice trees that have ordinarily been utilized in action acknowledgment, unsupervised realizing which has been utilized to distinguish irregular occasions, and fluffy rationale which is utilized to get larger amount data from states, for example, how the climate feels like. The last period of setting preparing comprises of setting conveyance, which should be possible by means of demand where explicit questions that are let go by the administration buyer, or distribute buy in model where buying in to a particular subject, for example, climate changes in London is being finished
This paper introduces the plan and execution of an arrangement of conveying setting mindful appraisals to understudies utilizing Samsung Smart looks for omnipresent learning situations. An exceptional part of the proposed framework is that it uses the MQTT as the key usage analogy and is coordinated with Google Classroom 
Architecture and Working
The general framework design is appeared in Fig. 1. As the figure appears, in the wake of enrolling and validation, an understudy wearing a savvy can buy in to educational modules point like “speed” in Physics and begin accepting evaluations identified with this subject on their watch. They will just get an evaluation dependent on their setting which can incorporate their area, for instance. As Fig. 1 shows, instructors or appraisal essayists can contribute their evaluations through a Web Portal. Publicly supporting can be utilized as such to assemble a vast appraisal bank. Creators can consolidate sensor esteems and setting triggers in their evaluations.
As Fig. 1 appears, the Context Processing Service (CPS) was executed utilizing Node.js. The understudy brilliant watches distribute their telemetry sensor information to the PONTE specialist. The Node.js server case asks the Google classroom APIs to get every accessible understudy and buy in to their fitting theme ids. The CPS running on a Node.js occasion is bought in to all the savvy telemetry information where this administration forms the information and matches it with reasonable accessible evaluations labeled with the asked for setting. The CPS has predefined setting types that can be prepared, for example, GPS or Heart rate information, for instance. CPS assesses the inquiry setting information against the telemetry information got and creates an inquiry. The subsequent in-situ questions are distributed back to PONTE where brilliant watches that are bought in to their focused on inquiry subject will in a split second presentation those inquiries to every understudy. The CPS likewise buys in to suitable evaluations, and as new appraisals end up accessible it forms the connection between settings inside evaluations against any telemetry information got. Since current Google class APIs don’t open an interface to refresh understudies’ reactions to questions, answer entries to the google classroom stage are presently being put away in a NSQL database. The particular inside arrangement for appraisals is gotten from QTI  and dependent on an IoT based Assessment Wiki  as appeared in Fig 2. As the figure appears, questions are spared in JSON arrange which incorporates setting mapping and dynamic inquiry and answer content.
The model has introduced a novel engineering for a pervasive evaluation framework dependent on the exceptionally adaptable MQTT convention. This framework enables instructors to create and distribute sensor-based evaluations fixing to explicit settings. Understudies, thus, can buy in to different settings including themes and sensors and get appraisals dependent on their present settings on a savvy. The framework has been actualized and is right now experiencing execution and power concentrates to guarantee that smart watch utilizes least assets, and that the framework will scale to hundreds or thousands of understudies.
 A. Labus, M. Milutinovic, ?. Stepanic, M. Stevanovic, and S. Milinovic, “Wearable technology in e-Education,” RUO. Revija za Univerzalno Odlicnost, vol. 4, no. 1, p. A39, 2015.
 (2017). Classroom API Google Developers. Available:
 (Issued1st September, 2015). IMS Question & Test Interoperability Overview Version 2.2. Available: w.html
 I. Zualkernan, M. Albayed, M. Al Solh, M. Tuffaha, and H. Al Muhallabi, “An assesement Wiki for Internet of Things,” EDULEARN14 Proceedings, pp. 2294-2302, 2014.