Voir la source de Smart-ring control for Smart-home ambience

{{Projet de spécialité étudiant
| titre= Smart-ring control for Smart-home ambience
| encadrants = [mailto:celine.coutrix@imag.fr Céline Coutrix]
| page principale = Projets de spécialité image}}


[[Catégorie:Projets de spécialité]]

== Students ==
*[mailto:son-tung.do@grenoble-inp.org Son Tung Do], M2 MoSIG.
*[mailto:henrik.helenius@grenoble-inp.org Henrik Helenius], M2 MoSIG.

== Context ==
=== Introduction ===
As a result of the explosive growth of the Internet of Things, everything is becoming connected. Furthermore, smart homes are more popular than ever. However, controlling the smart home can be a complex task, that has added complexity through complex user journeys in the smart home hub application on one's phone or on a tablet fixed on the wall of one's home. 

Problems arise as a result of bundling together various controls and properties to one user interface. As a result, tasks such as turning the lights on can now be unnecessarily complex and cumbersome, hidden behind several clicks. 

Because of the inconvenient controlling of some everyday-properties, we wanted to create a new way of interacting with a smart home. In more detail, this gave the user the control of the ''ambience'' of a given room. By ambience, we refer to four properties, which are: lighting, heating, window folds, and music. To ease the executing of the highly repetitive tasks relating to these properties, we propose Smart-ring control as a possible solution. 
=== Objective ===
The objective is to study whether a smart-ring control can have a positive impact on controlling the properties relating to the ambiance of the smart home when comparing it to the state-of-the-art, which is a smart hub application running on a smartphone. 

In the experiment, the user is tasked with controlling the properties of lighting (turning on or off), heating (increasing or decreasing), window folds (adjusting the position), and music (changing volume, play or pause, and skipping a song).


== Prototyping ==
=== Smart-ring ===
For the Smart-ring control, we designed and 3D printed a prototype where we have 2 rings fit in together, an Inner Ring and an Outer Ring. As shown in the picture below, the outer ring is a cylinder with a window gap in the middle, the inner ring is another cylinder and will be marked with different colors that indicate the active chosen property.

<center>
 {| class="wikitable alternance centre"
 |[[Fichier:Ring-prototype-1.jpg|300px|center|thumb|Ring Prototype left end]]
 |[[Fichier:Ring-prototype-2.jpg|300px|center|thumb|Ring Prototype right end]]
 |}
</center>
We design the smart-ring so that it will function as follows:
# Lighting: Rotation will change the value, binary control
# Heating: Rotation will change the value, continuous control
# Window folds: Rotation will open/close the window folds, continuous control
# Music volume: Rotation will change the value, continuous control
# Music selection, single-tap will result in toggling the play/pause state, and a double-tap will result in skipping to the next song (forward).

=== Smart-hub application ===
We designed the smart-hub application will functions as follows:
[[Fichier:Prototype-demo.mp4 ]]


== Experiment ==
=== Planning ===
We conducted the experiments in a Wizard-of-Oz type of manner, in other words by faking the interaction.

In this experiment, we will conduct the test with two prototypes consecutively. One test will be with the smart-ring (test 1), followed by a test with the smart-hub application (test 2). To balance the test, and avoid any repetition bias, half of the participants will complete test 1 first, while another half will do the test in the reverse order (smart-hub or test 2 first and then completing the smart-ring or test 1).

Before each test, the facilitator explained how to use the prototype as well as answered any possible questions. For the smart-ring test, the subject had a cheat sheet visible on the table which mapped the properties to the colors of the tracks on the ring.

For each test, we created an exact list of tasks for the test subject to execute. One task list was compiled for the smart-ring test and another one for the smart-hub application.
[[Fichier: Task-list.png |frame|center ]]

To collect data from the test subject themselves we used surveys that are also used in the industry. After each test with one prototype, we asked the participants to fill in a survey based on the System Usability Scale (SUS). Furthermore, at the end of the whole test session—in other words, after the subject had completed both test cases—we had another survey for more open form written feedback.

=== Setup ===
We conduct the experiment in a room where we have the following properties:
# Lighting: will be controlled by a switch on a wall
# Heating: will be a represented by a number on a screen (laptop monitor visible for the test subject)
# Window folds: will be electronically controlled by a switch located on a wall
# Music: will be played and controlled through a laptop in the room

We recorded the scene with the subjects completing each test case in the experiment and their interactions with the prototypes. This allowed us to analyze the various aspects of the experiment and the specific tests and the subjects after the subject had already left the room. Furthermore, it allowed us to collect better and more precise measurements. 
<center>
 {| class="wikitable alternance centre"
|[[Fichier:setup-scene-1.PNG|450px|thumb|center|Setup scene that we conducted the experiments]]
|[[Fichier:setup-scene-2.png|450px|thumb|center|The view of the camera we recorded video for analysis]]
|}
</center>

== Results ==
[[Fichier:Time-to-complete-each-task.png|800px|thumb|center]]
[[Fichier:Time-Difference-between-ring-and-hub.png|800px|thumb|center]]

[[Fichier:Screen Shot 2020-01-14 at 13.46.41.png |800px|thumb|center]]

== User Feedback ==
We gathered the written feedback survey regarding the ring prototype with participants at the end of the experiment and it can be summarized as follows:
* Pros
** The ring is light, simple and easy to control
** A nice unified control 
** Small enough to bring with you and always be on the user's finger

* Cons
** Users have to remember the mapping of the colors of the tracks and the property of the smart home
** There are 2 feedbacks mentioned that ring is not a favorable format and it's not easy to take on and off on demand

== Conclusion & Discussion ==
Based on the results, we can conclude that the smart ring that we proposed does not perform better than the smart-hub application.
There are several points that we want to discuss regarding the prototype and experiment that we conducted:

- All the participants are tested with the ring on their first look, training with the ring could make this interaction more familiar and users can get used to controlling it. Using a smartphone is however a very familiar device at this point to the whole general population, and therefore also to the subjects that we had in our experiment.

- The prototype production quality is not reflecting what an industrial grade smart ring would represent and does not have strong tactile feedback on changing the track, which leads people to people making mistakes. They had to really pay attention when choosing the track, the vision was however that users remember the track order and can control a property even without looking at the ring itself.

- The experiment itself had an inherit lag as the wizard-of-oz type mock interaction had a small—but noticeable—delay when controlling a property.

- The ring could be implemented with the haptic feedback so the interaction can be more fluid and also could reduce the size of the actual ring.

== References ==
Je, S., Rooney, B., Chan, L., & Bianchi, A. (2017, May). tactoRing: A Skin-Drag Discrete Display. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (pp. 3106-3114). ACM.

Han, T., Han, Q., Annett, M., Anderson, F., Huang, D. Y., & Yang, X. D. (2017, October). Frictio: Passive Kinesthetic Force Feedback for Smart Ring Output. In Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology (pp. 131-142). ACM.