Virtual Environments

Abstract

In the future, computers will be everywhere: carried by everyone and integrated into the environment. The increased computation and communication capabilities will enable environments to react to occupants through automated decision-making. On the path to intelligent environments, there are a number of hard problems. First and foremost, devices (sensors and actuators) must be made more approachable for average users. This dissertation aims to make applications for these intelligent environments easier to develop, thereby increasing the adoption and prevalence of emerging sensing and actuation technologies. Existing approaches to application development for these pervasive computing networks require detailed knowledge of low-level programming languages and embedded hardware. These current approaches also require adoption of new tools and languages that are not user-friendly. Intelligent environments will not be commonplace until average people can not only set up the necessary hardware and software but can also tailor the embedded software to personalized applications. This dissertation endeavors to create a software framework for pervasive computing networks that is easy to adopt and use. In resolving these challenges, this dissertation focuses first on programmers then extends to end-users. We start by standardizing device communication using highly available and widely accepted protocols. To ease the development effort for programmers, we allow complex sensor and actuator applications to be viewed as ii web applications. We term this work SEAP, or Sensor Enablement for the Average Programmer. SEAP makes available sensor data and communication capabilities that will allow average programmers to develop intelligent environment applications. However, to develop the massive amount of software necessary for personalized intelligent environments, it is also necessary to allow end-users to tailor the behavior of their sensors and actuators. This dissertation's second major contribution, Sensor Enablement for End-Users (SEEU) will provide a visual programming interface that will allow users to create personalized automated behaviors given available devices and data. This will require a principled design guided by end-users. Combining an intuitive interface and a middleware for sensor interactions, SEEU will be an end-user programming framework that will allow average people to create useful applications for their intelligent environments. For pervasive computing to truly bloom in these envisioned intelligent environments, it must be simple to develop new software for the plethora of potential available hardware. Upon completion, this dissertation will offer a software solution necessary to match the proliferation of hardware. The total package will usher in a new world of possibilities in pervasive computing.