Mobile sensing slowly became a part of our everyday life due to the evolution of the smartphone into a powerful sensing device. Popular consumer smartphones come equipped with multiple sensors to monitor a diverse range of human activities and commonly encountered contexts. Access to these sensors provides a particularly useful tool within ambulatory assessment and helps researchers understand the context and situation of the participant.

movisensXS provides researchers easy access to the Mobile Sensing features of smartphones. The additional insights derived through Mobile Sensing add context to questionnaire data, and can help determine the optimal times for interventions, proving especially useful for researchers using Interactive Ambulatory Assessment. Thanks to developments from movisens researchers can now easily access the benefits of mobile sensing.


Recommendations for the valid and reliable usage


Scenario 1: Assessment of smartphone usage

Smartphone usage provides a vast amount of information about a participant within any research study. Logging the participant's communication behaviour, their screen time, and app usage helps build a picture of their everyday usage. It can help researchers gain information about disturbances or interruptions of an activity.

Several functions help track the usage data, app usage, registering incoming and outgoing messages (SMS, calls, notifications), screen time or even collecting information about the music that's playing.


Scenario 2: Usage of integrated Smartphone Sensors

The sensors available within the smartphone can be used for different questions. For example, the internal acceleration sensor allows a rough estimate of physical activity.

Whilst Physical Activity cannot be measured accurately with a Smartphone, it can provide a very broad estimate of the amount of overall movement within a specified timeframe. Smartphone accelerometers lack the accuracy and uniformity of a dedicated activity tracker. To accurately measure activity please use a dedicated research quality activity sensor like the movisens Move 4.


Scenario 3: Position Tracking

The participant's location provides one of the most important pieces of context information.

movisensXS logs the location of a participant in an energy efficient way, balancing the need for accuracy with a careful regard to the impact on the smartphone battery. It uses a combination of GPS, WiFi and Cell tower signal to capture continuous location updates at a maximum rate of every 5 seconds and every 20 metres. The tracking transitions to a stationary state when the participant remains within 100 metres of a central position or no location update occurs for 120 seconds. Then tracking ceases until the participant leaves the 100 metre radius.

To track a participant within a building requires Bluetooth beacons. Placing multiple beacons at fixed points within a structure allows movisensXS to log them as nearby devices in an unsisens log file. It sends a bluetooth advertisement through iBeacon standard on supported devices and will protocol all other devices broadcasting with an iBeacon standard.


Scenario 4: Mobile Sensing for Interactive Ambulatory Assessment

The mobile sensing options provide additional triggering capabilities for questionnaires within an experience sampling study (interactive ambulatory assessment). For example, the use of the smartphone, including the use of selected applications or the location of the user can trigger a specific questionnaire or action.

The timing of the survey is precisely adapted to the respective research question, which increases the significance of the data. This precision allows fewer questionnaires throughout the day, and thus increases the overall study compliance.

A typical example is the Geofence option
This event fires upon a pariticpant entering, leaving or staying within a given geofence, allowing you to tailor location specific questions for your study.


Scenario 5: Mobile Sensing for Interactive Ambulatory Intervention

The mobile sensing capabilities also enable the precise delivery of interventions. Researchers can determine the optimal time for an outpatient intervention or to adapt the intervention to the situation.


Scenario 6: Social Sensing

Social interactions form the core of many experience sampling studies. Mobile sensing offers a variety of possibilities to track interactions both between participants and also between participants and other people. The information obtained by recording smartphone usage, such as communication behaviour (calls, SMS, chat app usage) also provides an insight in to the social world of the participant. The Nearby Technology detects whether a study participant is in the vicinity of another study participant. And the EAR app (Electronically Activated Recorder) allows the analysis of a person's conversations by recording audio snippets.


Example study: PEZ

The ZI (Zentralinstitut für Seelische Gesundheit) Mannheim conducts the "PEZ Study" on the effect of environmental factors on well-being and mental health. The PEZ study examines the relationships between environmental conditions, everyday stressors and personality traits in order to identify mechanisms that make people resistant or susceptible to mental illness over the long term. With the help of mobile devices, everyday movement patterns on satellite-supported maps are determined and combined with movisensXS smartphone questions to determine the state of mind in everyday life. Tissue sample analysis and imaging methods contribute to this large ambulatory assessment project.

The Mobile Sensing features featured in this project are location, activity type, steps and trigger when environmental parameters changed (noise, population density, vegetation).



Useful Information

"Mobile sensing uses the sensors of a mobile device (i.e. smartphone or tablet computer) to acquire data from the environment" Mobile Phone Sensing in Scientific Research; Bo Liu and A. Bulent Koc; 2015

Thus the simple use of a smartphone allows researchers to collect various information about the participants context. Furthermore, mobile sensing enables interventions to be optimally timed to fit the context.

movisensXS stores it's mobile sensing data in the Unisens format. Read more about this open source file format here.


Sensors in the Smartphone:



Limitations of Mobile Sensing

Along with the wide possibilities, mobile sensing also possesses some severe limitations due to Smartphone diversification, battery consumption and limitations of the operating system. This may result in more missing data especially when using participants own phones.

  • SIM-Card required
  • Android is very open
  • Battery Consumption
  • Android Updates (Privacy Improvements)

Available Data

  • Application Usage
  • Battery Level
  • Nearby Beacons (Through Walls, Requires Android 5)
  • Smartphone Running and Display On/Off
  • Location (Requires Internet Connection (SIM-Card))
  • Incoming Notifications (Requires Android 4.4)
  • Phone Calls (Must be Participants Smartphone)
  • SMS (Must be Participants Smartphone)
  • Physical Activity (Medium quality)
  • Steps
  • Music listened (Not compatible with all Android audio players)
  • Internet Usage measured by Data Traffic (Must be Participants Smartphone)
  • Electronically Activated Recorder (EAR) (Privacy Issues)

Literature

Mobile Phone Sensing in Scientific Research; Bo Liu and A. Bulent Koc; 2015.