MotionSynthesis Toolset

Body Sensor Networks and wearable computing devices are becoming more prevalent. They are being used for health monitoring, activity tracking, and fitness applications. Collecting the data necessary to develop the new concepts for these systems can be difficult. We present the MotionSynthesis Toolset (MoST) to alleviate some of the difficulty in data collection and algorithm development. This toolset allows researchers to generate a sequence of movements (i.e. a diary), synthesize a data stream using real sensor data, visualize, and validate the sequence of movements and data with video and waveforms.

Download

This tool is under constant development. For the latest version of all code and data please view our GitHub repository:

MoST on GitHub

For an interactive way to download data visit:

Interactive Database

Diary Generation Tool


Graph Panel

Graph Panel


The Graph Panel is used to load the file which lists the movements and their associated start and end postures. An example line from this file might look similar to this: <Sitting , Sit_to_stand , Standing>. The Graph Panel takes this information and displays a graphical representation of the connections between movements. Each movement is represented as a node in the graph, and selecting a node shows all directed edges associated with that movement (based on start and end postures). Each directed edge from a node represents a legal movement that can be made after the selected movement or movements.

The Graph Panel has several other functions. It can be used to add and delete nodes from the graph, and save a new movement file based on this user input. Current actions that can be performed in the Graph Panel include adding nodes, removing nodes, and showing the start and end postures. This is useful for growing the movement database that can be synthesized using the toolset, in an easy to understand visual representation. Without the use of this tool, it is difficult to visualize how each movement connects with the others.


Diary Generator

Diary Generation


The Diary Generator graphical user interface (GUI) is what is used to create the sequence of movements that will be synthesized. This tool includes ways to control the data synthesized, including a 'Best' or 'Random' option (as of now, Best selects the first trial of a subject rather than the best and Random selects a random subject and trial for the movement). There are also controls for movements that can be done for an extended period of time, in a 'Duration' menu option that allows users to select up to 300 seconds for the movement.

The panel on the right displays the current diary selections as they are made. This panel allows the user to keep track of the movements that have been selected as well as the options selected for each movement. For example, if a user selected Basic_Standing for 18 seconds it would look like Basic_Standing /B D-18.

This is what goes into the diary file when you click save diary, and is processed by the Data Synthesis MATLABĀ® tool.

Data Synthesis Tool


Data Synthesis


Once the diary is completed, the Data Synthesis tool can be started from MATLABĀ®. When launched, the Data Synthesis tool will request an input diary file. Once selected, the diary file is parsed to determine which specific movement data is necessary, and the order in which the data should be concatenated.

The Data Synthesis tool provides the primary outputs of the MoST toolset. These outputs are the sensor data files which represent the full stream of movements described in the diary file.

Data Visualization Tool


Visualization


The Data Visualization tool is the final tool in MoST. It is required that the Data Synthesis tool has been run on a diary before the Data Visualization tool can be run. When launched, the user will determine which diary output to view.

Data Visualization tool is broken up into two main areas, the data plot area and the video area. The data plot area loads the synthesized data files from the appropriate folder. Based on the sensor locations and sensor modalities selected in the Diary Generator, the relevant data will be displayed.

Acknowledgements


This work was supported in part by the National Science Foundation, under grants CNS-1150079 and CNS-1012975, and the TerraSwarm, one of six centers of STARnet, a Semiconductor Research Corporation program sponsored by MARCO and DARPA. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the funding organizations.