Mobile Usability Lab

User’s Manual

Version 0.1: March 2004

Table of Contents

1. Introduction	2
1.1 Background	2
1.2 Design Criteria and Specifications	2
1.3 Relation of User Manual to Protocol Manager	2
2. User Advance Preparation	3
3. Hardware Components and Assembly	4
3.1 Suitcase and Components	4
3.2 Color Quad Processor	4
3.3 Vanguard Pan/Tilt/Zoom Camera	4
3.4 XCam2 and WideEye Wireless Video Cameras	5
3.5 Wireless Video Receivers	5
3.6 Wireless Microphone Transmitters and Receivers	5
3.7 Camera Stands	6
4. Technical Data Collection	7
4.1 Running Protocol Manager Locally vs Remote Web Site	7
4.2 Using SVDA	7
4.3 Compression Codecs	7
4.4 File Naming Protocol	8
4.5 Troubleshooting	8
5. Technical Data Analysis	8
5.1 Using MVTA for Event-Based Ergonomic Analysis	8
5.2 Using Adobe Premiere-Based Tools for Video Editing	8
Appendices
A Photographs of Hardware Components
B Detailed Specification of Hardware Components

1. Introduction

1.1 Background

The Mobile Usability Lab (MU-Lab) is a tool that was developed by the Rehabilitation Engineering Research Center on Accessible Medical Instrumentation (RERC-AMI) for usability analysis of targeted medical devices by individuals with diverse abilities. The MU-Lab is an all inclusive research lab that has been designed for use in multiple environments. The system integrates data collection hardware and software, and complements a Protocol Manager that helps facilitate problem identification, planning, data collection and data analysis across the continuum of the medical instrumentation analysis process. It has been specifically designed for detailed multi-site product accessibility and usability analysis that can be conducted at a diversity of locations in the greater Milwaukee (based at Marquette University) and Bay Area (based at the University of California Ergonomics Lab).

1.2 Design Criteria and Specifications

The MU-Lab meets several design criteria that were established by the team. Overall, the system is portable, reliable, simple and intuitive and lightweight. Specifically, all the hardware components required for usability analysis fit into one standard carry-on suitcase, with the laptop stored in a separate case that can also be placed within the carry-on suitcase, if desired. All of the hardware components are securely housed within foam packaging so they are adequately protected during travel. The system allows for synchronized data collection of video, audio and sensor data to be used for usability analysis. Several models of wireless video cameras are included so you can select the most appropriate model for your research. Several of the wireless cameras can also be tethered because there may be situations where wireless communication is not permitted within the clinical setting. Audio collection occurs via a multi-channel wireless microphone system, which connects through the microphone port of the laptop. Several wireless sensor designs are also being implemented, including: contact sensors, accelerometers, and force sensors. A data acquisition card (16 channels analog, 8 channels of digital I/O) or LED video signaling are possible approaches to facilitate sensor data collection, although a National Instruments supported data acquisition card (NI-DAQ, PCMCIA port) and its associated drivers is one of the requirements for the data acquisition software (SVDA) to function properly. Real-time data collection is available with the use of Synchronized Video Data Acquistion (SVDA™) software (see Section 4.2). SVDA is used for video, audio and sensor data collection, and can facilitate video collection using a wide range of compression algorithms. The core tools for subsequent data analysis are Multimedia Video Task Analysis (MVTA™), MVTA for ergonomic task analysis and several video editing packages that make use of Adobe Premiere to add speed and a rich collection of features: Canopus DVStorm2 and Matrox RT.X100 (see also Section 5).

1.3 Relation of User Manual to Protocol Manager

A Protocol Manager (PM) is also included within the system design. The PM is a web-based software approach that helps guide the multi-site research team and on-site experimenters through all stages of the experimental procedure, including: medical device problem statement, pre-screening and tracking of subjects (patients and practitioners), activity performance observations during data collection, post-activity interview of participants, post-activity data analysis and comprehensive research documentation and data tracking. The entire data analysis process is coordinated through the PM, which is implemented in ASP.Net, C# and XML software in the Microsoft’s Video Studio .Net environment. The PM guides the user through all aspects of the overall task evaluation process for a given target category of medical instrumentation. This includes cases where the human subjects may involve a practitioner as well as the patient/client. The software also has several choices for navigation and saving, including a save and lock mechanism to protect data from being accidentally overwritten. The software can be implemented on a single computer without any Internet connection (where the data will be stored locally) or the data will be stored directly to the network database when an Internet connection is available. This is the topic of a separate technical report.

Figure 2.1: Suitcase Layout and Key Components of the MU-Lab

2. Advanced Preparation

Figure 2.1 displays a photograph of the MU-Lab carrying case with each hardware component clearly labeled. All of the hardware components that can be selected for inclusion in the MU-Lab case are listed in Table 1 below and a photograph of each component is included in Appendix A. The MU-Lab suitcase contains signage that can be used to verify that all items are in place. In special cases there may be a need for alternative cameras or camera stands, or selection of specialized sensors. In general, however, after doing a component check, the key task is to check battery charges.

3. Hardware Components and Assembly

A general description of the hardware components and instructions for use are included below. For more detailed specifications of the system components, including model numbers and manufacturer information, please refer to Appendix B.

3.1 Suitcase and Components

All of the MU-Lab hardware components fit within a standard carry-on suitcase (approximately 9”x14”x22”), as shown in Figure 2.1. The suitcase is protected with hard plastic shielding and contains a customized foam insert to permit safe and secure travel with the MU-Lab. All of the hardware components have specific locations within the suitcase that are clearly labeled. There are also several pockets located within the case (on the top side that opens) that house the cables and remote controls available for use. There is also has a zipper feature to expand the top section of the suitcase to increase the depth by about 3”, for extra equipment if desired. A laminated copy of the Quick Start guide, System Diagram and Pocket Inventory are included for your reference (hardcopy in the suitcase and electronic copy in “Instructions” folder on the laptop.

3.2 Color Quad Processor

Since SVDA currently supports only one video input for data collection, the Grand Magic Guard II color quad processor is used to capture several video streams and compile them into one output. The unit supports input from up to four separate video cameras (with standard RCA video cables) and outputs one real-time composite video that is inputted into SVDA (with an RCA to USB adapter cable). The unit is a purely Plug & Display hardware component, so no software is required for use.

The quad processor must be powered via a wall plug with a DC adapter that plugs into the DC IN port on the back of the unit. There is also a VIDEO OUT channel located on the back of the unit that should be connected with an RCA cable to the RCA to USB converter and then into the USB port of the computer. There are four video channel connections, labeled VIDEO IN1 through VIDEO IN4, which are located on the side of the unit for input from up to 4 video cameras (RCA cables). Once all of the RCA cables (both into the quad processor and out to the laptop) have been connected with the appropriate cables, you can turn on the power with the switch that is located on the side of the unit.

There are several display options with the quad processor that you can be switch between, depending on the number of cameras in use and your requirements. You can display any single video feed, any two video feeds via Picture in Picture (PIP), or up to four video feeds with a quadrant display It is recommended that you initially display one field of view from a single camera while setting up the cameras in your research environment so the details of interest are captured within your videos. You can use either the remote control or the buttons located on the front of the quad display to program and switch between available options. You can also display the time, date and/or separate channel names, which must be programmed with the remote control once the unit is powered.

3.3 Vanguard Pan/Tilt/Zoom Camera

The Vanguard Pan/Tilt/Zoom camera can be used either wirelessly (with the video receiver plugged into the quad processor) or tethered with a standard RCA video cable (plugged directly from the camera into the quad processor). An ACDC adapter and power supply are required for the camera to operate, which should be plugged into the back of the camera and a standard wall supply. A remote control is used to adjust the camera settings and to pan, tilt and zoom, which has clearly labeled buttons. There are left, right, up and down buttons to control the panning and tilting positions of the camera as well as a center button that moves the camera back to its home starting position. There are also buttons to Zoom In and Zoom Out as well as an Iris button to adjust the amount of light that enters the lens. There is also an Autofocus button and Manual Focus buttons to allow you to adjust the focus of any particular area of interest within the camera’s field of view.

This camera is equipped with the standard tripod mounting screw threading, so it can be easily attached to any tripod. In some situations you may want to mount the camera upside down, so the picture should be flipped so it is displayed in the proper orientation. The following commands should be used to flip the image when the camera is placed upside-down:

1. Press and release the C1 button.
2. Press and hold the Setup button on the remote until the green C1 button lights.
3. Press the Iris Up button to flip the picture.
4. Press the Setup button again to save this setting.

3.4 XCam2 and WideEye Wireless Video Cameras

The XCam2 and WideEye cameras have the same basic components but different lenses. These cameras can be powered with either a battery pack accessory so they are wireless, or with an ACDC adapter and wall supply. If the battery pack accessory is used, the batteries will supply power to the camera for several hours of video collection. It is important to make sure the camera and receiver are tuned to the same frequency and the antennas adjusted appropriately to provide the best signal quality.

3.5 Wireless Video Receivers

If you wish to transmit the video signals wirelessly, a separate video receiver must be used for each camera. Each video receiver requires an ACDC adapter and power supply for operation. There is a power switch located on the side of the unit a red LED on the front of the receiver to indicate when the unit is on. The appropriate number of receivers should be plugged into one of the four available input channels of the quad processor using a standard RCA video cable.

It is also necessary to tune each camera and corresponding receiver to the same channel, which is different from all the other camera channels being used. The receivers have a switch for channel selection (A, B, C or D) that is located on the left side of the unit and the cameras have a switch for channel selection, with the location dependent on the camera model. The Vanguard Pan/Tilt/Zoom camera has a channel selection switch on the bottom of the unit and both models of the XCam2 cameras have switches on the top of the unit, underneath the black rubber insert. With the XCam2 models, the switch is not clearly labeled but channel A is located closest to the camera’s antenna.

The receivers can detect a signal from a camera that is located up to 100 feet away, even through walls since the signal is radio frequency (RF). Once the camera and receiver have been placed in their desired locations within the experimental environment, you should adjust the antennas so they face each other. The side of the camera antenna with the dot should face the receiver antenna with the four squares and the positions should be adjusted until optimal picture quality is reached.

3.6 Wireless Microphone Transmitters and Receiver

A discrete two-channel receiver and two wireless microphone transmitters are included to facilitate audio collection during your research. The normal protocol is to use the two wireless transmitters to capture and record the audio of the researcher and subject during experimentation. Since only two transmitters are included within the MU-Lab, some situations may require you to switch the available microphones between multiple subjects. For the best performance when using both microphones at the same time, it is recommended that you maintain a distance of at least ten feet between the transmitters and receiver. It is also recommended that the receiver antennas be positioned first straight up, and then angled out slightly so the angle is never more than 45 degrees from vertical.

Each microphone transmitter has a power/frequency switch, which is located on the top of the unit. This switch is used to turn the system on and off as well as to select an operating frequency for the transmitter. When the microphone transmitter is switched to F1, the LED will light up, which indicates the battery is good and the transmitter is functioning properly. It is also necessary to turn on the receiver power, either one or both channels depending on the number of microphones in use. The microphone receiver should be plugged into the microphone port of the computer, prior to starting SVDA.

The microphone system operates in the 169-172 MHz bandwidth, where interference is minimal. Either of two transmission frequencies can be selected: F1 (169.445 MHz) or F2 (170.245 MHz). It is suggested that you check the amount of interference in your research environment before you start collecting data. To test if there is RF interference in your research area you should:

1. Make sure the microphone transmitters are off.
2. Turn both of the microphone receiver switches (F1 & F2) on.
3. Both LED’s should turn red and if either or both of them turn green then there is RF interference in the area, which may affect the quality and performance of the microphone unit.

3.7 Camera Stands

There are several camera stands that can be included within the MU-Lab. All of the cameras and stands are equipped with the fittings necessary to secure the cameras to the stands. You can select up to 3 camera stands for inclusion in the carrying case, although several of the tripods have separate bags that can be used if you require more models than will fit in the case. The height range for each camera stand is listed in Table 2 below.

4. Technical Data Collection

This section targets technical data collection, i.e. data collection by video, audio and sensors. Subject questionnaires and the recording of observations during experimentation are made using the Protocol Manager, which is described in a separate technical report.

4.1 Running Protocol Manager Locally vs Remote Web Site

The Protocol Manager is available to guide you through the necessary procedures to adequately carry out your research protocol. The Protocol Manager software will be installed locally (as XML files) onto the MU-Lab computer and is also available on the internet at www.rerc-ami.org/D1PM/UserLogin.aspx.

If you have an internet connection or wireless capability in your research environment then you can access the Protocol Manager directly on the internet, so the data that is inputted and collected during your research session can be directly stored on the online server. When the information is stored locally the MU-Lab computer acts as the server until the data is uploaded to the online server. Otherwise, the MU-Lab computer acts as a browser over the internet and the information is sent directly to the server. If you are not online while filling out the subject and research protocol forms then you will store the data locally and then upload it onto the server at a later time.

4.2 Using SVDA

Open SVDA on your laptop. SVDA is automatically set to record data and no video, to change this look for the two scroll menus on the bottom called “Data Mode” and “Video Mode.” Select preview mode if you only want to preview the data and/or video without creating files. The file names can not be changed until after they have been collected. File names typically look like “data040106_110176_538.***” where the stars are either BDF for data files or AVI for movie files. Now go to the “Video” menu at the top of the program window and choose “Video Settings.” If an error comes up saying something about an “access violation at address….”, click ok and close down the program, unplug the USB video connection and plug it back, and then restart SVDA.

Once the “Video Settings” screen appears correctly there should be some values and codecs already selected, these can be easily changed. On the top right is the option of choosing to turn on a video compression mode. Selecting “no” will leave the video as a full .avi file and there will be no compression (approx. 600MB/min). Selecting “on the fly” means that SVDA will access whatever video codec you have selected (in the scroll down box located below the “compress Mode box”). This will cause your preview window of the video/data to only refresh once every couple of seconds, but it is not affecting the quality of collection. Selecting “after capture” will cause the video to be initially captured as a raw DV.avi file; once capturing is complete SVDA will use computational resources to encode the video to the selected format. This process of encoding after capture can take up to 3-4 times the length to encode as the movie length that was captured. Thus, a 30 second video capture can take up to 2 min to encode. Also keep in mind that while capturing video SVDA is writing a raw DV file to the hard drive, which it later erases, but is usually at a rate 600MB/min.

Next to the compression mode is the compression type. This box will let you select what you want to compress: video, audio or video+audio. Below the compress mode box is the scroll list to choose the “video compressor.” This list is generated by SVDA accessing the OS list of video codecs available on the computer. To add more codecs they must be installed, the OS must recognize them, SVDA restarted, and they should then show up in this list. Immediately below the codecs are two values, “key fr. rate” and “video quality.” They both are automatically set to -1 by SVDA. This means that the codec being used will compress at it’s “natural” setting. “key fr. rate” refers to how many frames per second you want to be key frames. A key frame is a new image taken from the CCD of the camera which is then compressed into the video file, so if a value of 30 is put in, and you are collecting at 30fps (located to the right of the video compressor box) then every frame should be a key frame. The “video quality” value usually is a percentage of compression strength/weakness, with a range of 1-100. Where 1 would be very little compression and 100 would be utilizing the full strength of the compression algorithm. From experimentation there is little change in the video quality when using a value of 1, 25, 50, 75 & 100. It is recommended to leave the “video quality” setting at -1.

4.3 File Naming Protocol

A standard format has been established for naming and storing the various files that will be produced during your research. A separate folder should be created for each subject on the MU-Lab computer, where all the files associated with that subject should be stored. Each subject is identified by a unique numeric label, which will be used for all research and discussion purposes. The assignment of this numeric label is coordinated by the Protocol Manager, with the Marquette series beginning with 1000 and Berkeley with 2000. The following is the format for naming the raw data files:

1. PAT/PRV label for patient or provider identification
2. Subject Identifier (four digits)
3. Equipment identifier (2 letters)
4. Session identifier (2 digits)
5. Task identifier (1 letter)
6. Trial identifier (1 digit)

For example: PAT1001ab11a1 or PRV2001ab11a1

4.4 Troubleshooting

• If you receive an error from SVDA: restart the program and/or the computer after checking the USB video input connection and making sure all of the hardware is properly connected and powered.
• If there is significant interference with the wireless cameras: check the battery power, make sure the camera/receiver pair is tuned to the same channel, relocate one or all of the receivers so they are as far away from each other as possible or realign the camera/receiver antenna pair(s).

• Make sure to look for power switches on the hardware devices (cameras, quad processor, microphone).

4.5 Compression Codecs

For high quality video, use the Microsoft Video codec or the new MPEG-4 DIV codec. For low quality video, use the Cinepak or Div codec. For both use the during capture setting and -1 for the key fr. rate and video quality setting.

5. Technical Data Analysis

5.1 Using MVTA for Event-Based Ergonomic Analysis

MVTA is used for post-processing data and task analysis procedures, specifically time and motion analyses of video recorded activities. In general, MVTA is used to identify events with terminal break points during a timed activity for usability analysis. Video and sensor data collection is synchronized so it can all be viewed on the same timeline during video playback. It is also possible to replay the video file at any speed, including real-time, slow motion, fast motion or frame by frame in both the forward or reverse direction. There are other features that allow the researcher to control features such as replaying an event in a continuous loop or displaying an arbitrary event or point in time. MVTA can also produce conventional time study reports and the frequency of occurrence of any particular event, which may be helpful in usability analyses.

The main program window appears when MVTA is started and you must choose between two analysis types: Task Analysis or Data Analysis. The Task Analysis option is for viewing and analyzing video data and the Data Analysis option is for viewing and analyzing sensor data. For a more detailed discussion of the MVTA software, please refer to the MVTA User’s Manual, which is available for reference.

5.2 Using Adobe Premiere-Based Tools for Video Editing

Adobe Premiere is a valuable package that adds video editing capabilities, and can be used digitally edit and enhance video files. In addition, sophisticated video editing packages (Canopus DVStorm2 and Matrox RT.X100) consisting of video accelerator boards and associated third-party software that make use of Adobe Premiere add speed and a additional collections of features.