Physiological Signals

5.1.4. Output

Applications on mobile devices have a great variety, ranging from applications like games to more performance oriented office applications. The state of the art analyses showed, that it is possible to determine different emotional, affective or cognitive aspects of an user from physiological signals. As the presented examples in the state-of-the art section in chapter 3 show, games and applications, which have the goal to entertain an user, may benefit more from the affective state instead of cognitive aspects. On the other hand ap-plications, aiming for a certain performance level of the user and the goal to keep an user within this level, may benefit from information about cognitive aspects. Examples for such applications are learning and office applications or applications that need a high amount of mental resources.

Based on that analysis the concept will cover affective and cognitive aspects to allow support for a broad variety of applications.

5.2. Physiological Signals

For further thoughts about the concept and the model it is important to decide, which physiological signals are measurable in mobile scenarios. A broad range of signals is used in different applications, but not all of these are applicable for mobile scenarios. Require-ments have been described in section 5.1.1. In this section, different physiological signals will be reflected briefly regarding these requirements. A decision will be made about the signals as well as sensors will be presented, which will be used.

5.2.1. Overview of Physiological Signals

Not all physiological signals can be measured without impairment of the user. For the use in mobile settings, it is important to use unobtrusive sensors that allow wireless transmis-sion of data as described in the requirements section 5.1.1. The sensors need to be as small as possible to achieve a high user acceptance.

For that reason, the concept is based on electrodermal activity and heart rate. Both signals can be measured with relatively small sensors. As described in 2.5.1, electrodermal activity can be used to determine the arousal level of an user. Heart rate can be used to determine the valence in some cases or to run a spectral analysis for determination of mental effort (see chapter 2.4).

Other physiological measures like electroencephalography (EEG) or electromyogram (EMG) require at the moment relatively big sensors or sensors in a prominent place. For example EMG electrodes are commonly placed in the face. EMG is used in many research projects and applications, described in the State of the Art section in chapter 3. But most

of these projects had not to deal with mobile scenarios and its requirements.

Due to these constraints, a heart rate monitor that allows Bluetooth transmission as used in sports is used for heart rate measurement (e.g. Polar or Zephyr HxM). For measurement of electrodermal activity, the Q-Sensor from Affectiva will be used as input channel in the model. In early studies, wired stationary devices were used for first evaluation of applications. In later stages, Q-Sensor, Zephyr HxM and Polar H6 were integrated.

Two heart rate sensors were integrated because of the different options and additional sensors they offer. Polar H6 offers a higher battery lifetime but otherwise Zephyr HxM offers more additional information measured by the heart rate monitor. Depending on the application scenario, one of the sensors might be more suitable than the other. Q-Sensor was chosen because it measures wirelessly and reliable Electrodermal activity (EDA).

5.2.2. Affectiva Q-Sensor 2.0

The Q-Sensor was originally developed at the Massachusetts Institute of Technology by Poh et al. [PSP10] and was distributed as Q-Sensor 2.0 curve by Affectiva. Besides skin conductance it measures temperature and movement. It allows wireless transmission of the measured data via an integrated Bluetooth module.

Figure 5.1.: Q-Sensor 2.0 from Affectiva [Aff11]

As shown in figure 5.1 the sensor is worn as a wristband. Two Ag/AgCl electrodes are placed at the inner side of the arm. As described in the background area, EDA is normally measured at the palm. A study described in [PSP10] shows that EDA can also be measured with the Q-Sensor at the distal forearm. Furthermore the participants of this study did not feel discomfort in a long time study. In this long time study of Poh et al.

[PSP10] the Q-Sensor was also tested against devices, which are approved by the Food and Drug Administration (FDA) in the United States. Results showed, that the Q-Sensor offers very high accuracy comparable to FDA approved devices.

5.2. Physiological Signals

With help of the integrated sensor for movement of the arm and body temperature, influencing factors can be controlled in the interpretation of EDA values. Q-Sensor has been used in several other research projects, e.g. in the evaluation of child-robot interaction [LHMP13].

5.2.3. Zephyr HxM

Zephyr HxM (shown in figure 5.2) is a heart rate monitor consisting of a strap and a trans-mitter. Transmission is done wirelessly via Bluetooth. Out of the cardiovascular measures, Zephyr HxM measures heart rate and RR-intervals. Furthermore calories, steps, speed and distance can be measured with integrated sensors. Integrated algorithms address the problem of noise and movement artifacts [Tec10].

Zephyr HxM has been used in different research projects like in a phone-based health assistant [SBVL11], the involvement of audience at public events [PTS⁺10] and different research projects at the National Aeronautics and Space Administration (NASA) [RC12].

Figure 5.2.: Zephyr HxM BT [Tec10]

The data is transmitted wirelessly via Bluetooth 2.0. A version supporting the Bluetooth Smart standard was not available at the time of this work. Because of the Bluetooth 2.0 transmission, the battery life time is only about 26 hours. In comparison to Polar H6, bat-tery can be recharged with an included charging station. An open software development kit is offered by Zephyr Technologies and the heart rate monitor is supported by different publicly available applications.

5.2.4. Polar H6 Heart Rate Monitor

The Polar H6 (shown in figure 5.3) heart rate monitor consists of an adjustable strap and a transmitter. The electrodes are on the inside of the strap. Strap and transmitter are both water resistant and have a battery lifetime of around 300 hours. The Polar H6 measures RR-intervals, heart rate (HR) and several other additional information.

[NDJ⁺09] proved the validity and reliability of former versions of the H6, the Polar

RS810 in comparison to electrocardiogram (ECG) devices. The results showed, that short term Heart Rate Variability (HRV) measurement is as reliable and valid as stationary de-vices. Several commercial applications exist, that support the Polar H6.

Figure 5.3.: Polar H6 [Ele13]

The H6 supports the Bluetooth smart standard, which allows the wireless transmission with low energy consumption leading to longer battery lifetimes. The transmission range is approximately 10 meters, which should be sufficient for most scenarios.