IoT Presence, Sensing, and Awareness in the Home
Project overview
The problem
We are already surrounded by ‘Internet of Things’ devices and they are becoming more common in homes, workplaces and public spaces. But how much do we know the kind of data these devices are sensing, storing and uploading to databases? Frequently, IoT devices do not give users feedback about what they are sensing and recording. Even if they do, it is hard to notice. On the other hand, we usually do not want to be constantly disrupted by our devices. Most users prefer their technology to dissolve into the background. This presents a challenge of achieving a balance between this two.
Our work
Our research explores this paradox: investigates the design of an IoT table that operationalizes these ideas to make a device that allows people to be aware of IoT sensing without demand a ton of their attention.
My role
Interaction design, prototype, experimental procedures design, laboratory studies, qualitative data analysis, writing paper.
Advisor
Teammate
Michal Luria (PhD Candidate at CMU HCII), Benjamin Stone, Ashvik Awasti
Timeline
Jan - Aug 2018
Introduction
People are increasingly living in spaces augmented with IoT technology meant to improve their lives; some examples include Amazon Alexa and Apple Watch. One of the first to envision interaction with ubiquitous computing that blends into the background is Mark Weiser, in his vision of ‘invisible computing.’
However, the problem is that ‘invisible’ IoT devices lack transparency, which frequently raises concerns among their users as they are not being informed of what their devices are sensing and perceiving. While company owners of such devices state that only saying an “activation word” results in an active device, news items have reported several instances where devices such as Amazon Echo were ‘listening’ to a conversation without the users’ knowledge. The most common practice to address this issue is marking spaces with physical signs that inform the area is under surveillance. The signs are detached from the individuals’ interaction with the device. It was our idea that perhaps continuous feedback from an environment or a device inherent in the interaction could add the needed transparency.
In order to test this idea, we designed a coffee table to investigate how IoT technology might be more transparent to users without making computing explicitly more visible. In coffee table’s design, we implement the notions of ambient displays and social presence to allow the table to communicate what it is sensing and perceiving (processing) data from its surroundings.
We explored the feedback and interaction patterns for the coffee table by asking potential users to engage in various everyday scenarios in its presence and by observing their reactions and behaviors around it. The evaluation of the table raised many interesting issues and questions about invisibility, transparency, and control. We discussed the themes that emerged and their implications for the design of the next generation of IoT technology.
Design
The table prototype
We created an interactive coffee table designed with the goal of communicating social presence and presenting information to participants about the sensing in their space.
Feedback and control
We set out to design feedback for the table that would give participants in the space a sense of the table’s spacial awareness and recording status. Focusing on a few common sensing capabilities of IoT devices, we wanted the table to acknowledge when people are in the space, when someone is talking, and when someone is moving.
In order to do so, we created several animations that reflected such occurrences. The animations were a combination of automatic sensing and manual control using the “Wizard of Oz (Woz)” method.
For the automatic sensing and feedback generation, a microphone was attached to the center shelf of the table (hidden with drawers), and automatically captures the volume in the room. This allowed us to accurately sense and present the volume level when someone in the room spoke. To present the number of people in the room, their movement and the instances in which they spoke, we used WoZ manual control.
Animations
For the automatic sensing and feedback generation, a microphone was attached to the center shelf of the table (hidden with drawers), and automatically captures the volume in the room. This allowed us to accurately sense and present the volume level when someone in the room spoke. To present the number of people in the room, their movement and the instances in which they spoke, we used WoZ manual control.
Each rock represents a person, fish swim towards the speaking/moving person; the brightness and radius of the circle will change according to the speaking person.
Each flower represent a person; the flowers' rotation speed represent people's talking and their talking speed.
Each bright circle represent a person; the wave moves outwards the speaking person.
Each rectangle represents a person; the brightness and length of the rectangle changes according to the speaking person.
Evaluation
We evaluated the interactive table and feedback for IoT devices’ sensing and perceiving by asking participants to play out various scenarios that depicted everyday activities. The goal was to define a procedure that would generate responses to having the interactive table present in daily activity, as well as activities that include some concern for privacy.
We recruited 25 participants for this study. Since it's an exploratory study, we iteratively modify our experiments (scenarios, procedures and animations) according to participants' behavior.
Result
Coming soon…
Future work
Coming soon…
