Introduction
Most applications of consumer IoT systems either operate on the explicit command of the user or are fully automated. Commanding an IoT system requires a user to spend a substantial amount of attentional demand on the system. On the other side, fully automated systems can have a high level of complexity, lack sufficient reliability, do not fit current and changing lifestyles of users and might create the sensation of loss of control .
Interactive Intentional Programming (IIP) recognizes these problems and proposes a framework to capture the intents and preferences of the end-user to make suitable actuations. The researchers acknowledge two main areas of improvement: better methods for capturing scenarios, intentions, and preferences, and second, the creation of a feedback loop to facilitate
adoption and learning over time.
Such a feedback loop can become complicated. The intervention user interface allows the user to intervene with the automated behavior of the system. This allows the systems to operate at high automation levels while only involving the user in the loop when necessary. The implementation of intervention interfaces have not been researched previously.
This research applies the Intervention User Interface principle on an IoT system that uses the basic concepts of IIP. The system controls a research office environment that strives to have a high level of automation and autonomy while still giving the user the sense of control. The result is “PII”, a Peripheral Intervention Interface for IoT.
Prototype & User Study
The variations for PII were realized in Processing programming environment. The laptop, in the picture on the right, was used to run the Processing sketch and to control the autonomous behavior of the system (initiated by the researcher). The graphical user interface was casted on a smart phone that was docked in the PII prototype which was placed in the working environment of the participant. Allowing the participant to interact with the actuators in the environment. The dock enabled the user to interact with the interface that was running on the laptop. The dock contained a big space bar button that triggers an intervention, the intervention button. Making it easy and quick to use. Adjusting parameters in the Processing sketch would send data to a Teensy that controlled different actuators in the environment. Light was emitted by multiple ledstrips that were positioned next to the participant. A heating mat was used to give the participants a more direct feedback about temperature changes. This was necessary due to the fairly short testing period, i.e., the user needs quick feedback to be able to detect change in such short periods of time. The heating mat was placed directly on the table surface, which also functioned as the working area for the participant. Music was played with the use of a Bluetooth speaker which was placed directly behind the small office divider, out of sight for the participant.
Video Explanation
Research Framing
Different levels of information are applied to three interfaces to examine the level of attentional demand and involvement preferred by the user (i.e. until what extend does the want to be involved in the loop). What level of attentional demand (explicit, peripheral, implicit) is preferred by users when interacting with an intervention user interface in the Internet of Things smart office?
Conclusion
This research shows first implementations of the Intervention User Interface principle. The research focused on the preferred attentional demand by users and how this affects their sense of control. The user study has shown that users prefer highly informative interfaces. This helps them with feeling reassured and letting go control of the system. Different levels of information or attentional demand did not have a significant effect on the user sense of control. In addition, this research gathered a set of design principles that can be used by researchers and practitioners for further research on or implementation of Intervention User Interfaces.
The design principles:
- Only inform users with the essential automated decisions.
- Use the IoT environment to inform the user about the systems state in non-intrusive ways.
- Allow the user to access additional information if necessary.
- Give direct explicit feedback for interactions initiated by the user.
- Do not invite for interaction. However, keep the barrier to do so low.
- Give users the same controls the system has.
Want to know more?
Contact me!
I would love to tell you more about the details and results.