Main Page

Overview

This page contains information about the on-going projects in human activity recognition at the University of Washington.

Platforms

The current platform that we're using consists of an embedded processor component, called an iMote2, and a sensing component called the MSB2 (Multi-modal Sensor Board). The combination of the iMote2, MSB2, and the system software is often referred to as the MSP (Mobile Sensing Platform).

In addition to the MSP we've also been using a number of cell phone platforms to provide wide area communications, display, local interaction capabilities, and additional sensors.

Component Details

More information about the various platform components can be found here:
• iMote2 - The Linux embedded computational platform that forms the heart of our wearable sensing platform
  • More information about the iMote2 and the Linux software can be found at the PlatformX wiki
• Sensor Boards - A brief description of the sensor boards that are being used

Software Tools

Logging Software

• msb server - the program used to record data from the sensor board

UWAR Files

UWAR files (University of Washington - Activity Recognition) are a simple binary format used to store data.
• overview - a brief description of the uwar format
• uwar tools - contains a description of the various tools created to work with uwar files

Matlab

• Plotting wide images - Program to plot images with an extremely wide aspect ratio (i.e. hours of data plot vs. a small y-axis with good time resolution)
• Compositing Images - Matlab function to composite images (like those generated by plotting wide images.

Wiki Help

Please see documentation on customizing the interface and the User's Guide for usage and configuration help...

Wikipedia has some helpful information on how to edit wiki pages:

• MediaWiki FAQ
• Cheat Sheet PDF version
• How to edit a page
• Syle Manual
• Tutorial on using Pictures in a Wiki
• How to use lists in the Wiki

Wiki Add-Ons Used

• GeSHi has been installed along with an extension for MediaWiki. This enables code to be highlighted in several different languages.