Related Work
Activity detection projects	Many research projects have focused on detecting activity in many environments. One project that is related to this work is Oliver and Horvitz's work (2002) on the SEER system which detects activity in the office using layered HMMs. Their layered method of detecting activity was able to achieve an almost perfect detection of activities (99.7% accuracy) using similar sensors as my project. Although, my project does not approach the same accuracy as the SEER system, my work extends the work of Oliver and Horvitz by applying the detection of activities. In my project, I explore the feasibility of using activity detection to assess productivity of office workers. Yang et al. (1997) used HMMs to recognize, characterize, and emulate simple human skills. They use HMMs as their learning algorithm and suggested applying it to improving robotic control. In this project, we aim to apply detection of activity to inform them of their productivity.
Interruptibility vs. productivity	Hudson et al. (2003) have done research on predicting interruptibility of office workers. Interruptibility and productivity may be negatively correlated. For example, interruptions may reduce the productivity of a worker that is not interruptible. However, these two terms are different. First, productivity have many aspects and interruptibility is one of them as I mentioned above. Other aspects of productivity include one's motivation to work, the quality of the task being performed, the susceptibility of the self to distractions from work, etc.
Awareness	Begole et al. (2003) tracked people's availability using the Awarenex system and used the information to increase awareness of people's patterns within a group. In this project, we track people's activities and the aim is to use the information to inform the person that is being tracked about their own productivity.