SymbIoT

Designing spatial interactions for the Internet of Things.

Overview

SymbIoT is a long-term research project about giving end users the ability to program environments filled networked objects in the Internet of Things (IoT). It consists of programmable network nodes, an HTTP API for controlling the nodes, and a user interface for graphically programming the nodes with respect to the user's environment. The primary focus of the SymbIoT architecture is the user interface.

I was supported by a Summer Undergraduate Research Fellowship (SURF) from UC Berkeley while working on this project.

Background and Research Question

With the rapid increase of network-connected devices and things, users will need a way of programming them that scales well. Currently, approaches to programming the IoT are object-centric--we are used to programming individual devices rather than the overall interaction between devices. SymbIoT experiments with new ways that allow users to focus on designing coarse-grain setups of IoT objects, rather than on fine-grain device configurations.

First Iteration: Hardware and Command Line Interface

First, I built the hardware nodes that receive commands from a central controller. Each node is an Arduino Uno microcontroller connected to both a I2C BlinkM programmable LED and a Nordic RF24+ radio chip. In C, I programmed the Arduino to poll the radio, transmit and process bytes, and illuminate the BlinkM led with various brightness and color according to a low-level domain-specifc language. For example, to set node 2 to 100% brightess, from the central controller, one would enter: >>> s2,100

Second Iteration: Web API

I wrote an API that translates HTTP requests into firmware-level commands for the nodes. I used Express.js running on a Node.js server, integrated with MongoDB for keeping system state. This step was important to be able to control the nodes from a web browser.

More explanation might go here.

Third Iteration: User Interface

Designing the software user interface was the principle challenge of the project, answering the fundamental question: how should people program IoT object-rich environments, a task that has never had a UI designed for it before?

To explore this question, I first conducted a pilot study to test the domain of real-time direct manipulation user interfaces. Two trends I noticed were:

I've mocked up new UIs to test, but these plans are currently on hold.