A Stress-Detection System Based on Physiological Signals and Fuzzy Logic
Summary (5 min read)
II. LITERATURE REVIEW
- Stress detection has been considered from different points of view and approaches.
- Considering this, there exist many previous works related to this topic.
- Continuing on this research line, the research provided in [8], [9] proposes a system gathering FT, GSR, and blood volume pulse (BVP).
- The main purpose of this approach is to recognize emotions, interest, and attention from emotion recognition.
- Moreover, the work presented by Healey and Picard [2] deserves special mention, since they are considered to be pioneers on stress-detection field.
III. PHYSIOLOGICAL SIGNALS
- Several indicators have been considered within the literature to detect stress (Section II).
- In other words, extracting skin conductivity requires a small current passing through the skin.
- The main parameters of GSR, such as basis threshold, peaks, or frequency variation, vary enormously among different individuals, and thus, no general features can be extracted from GSR signals.
- Concretely, a stressing stimulus provokes a similar reaction, increasing the number of heartbeats when facing these situations.
- Nevertheless, HR signal behaves in a different manner depending on the stimulus and the individual.
IV. DATABASE ACQUISITION
- Every biometric system requires a set of different biometric acquisitions to train, validate, and assess the whole system [22].
- This section provides an overview of how the data set was built considering the experimental setup and the characteristics of the database.
A. Overview
- The experiments were carried out in the Faraday room in the Human Psychology Laboratory, Psychology Faculty, Complutense University of Madrid (UCM), endowed with electromagnetic, thermal, and acoustic insulation.
- The main aim of this step is to collect HR and GSR signals from each participant.
- The device proposed to carry out these experiments is I-330-C2 PHYSIOLAB (J &J Engineering), which is able to process and store six channels, including electromyography, ECG, respiration rate, HR, and GSR.
- Sensors were attached to the hand’s right (or left, but not both) fingers [23], wrist, and ankle in order to acquire both HR and GSR, avoiding sensor detachments, unplugged connectors to the analog-to-digital converter, and/or software acquisition errors.
- Moreover, the sample acquisition rate is made with one sample per second for both HR and GSR.
B. Participants
- The participants were mainly students from the Psychology Faculty (UCM) and Social Work (UCM), with a total of 80 female individuals of ages from 19 to 32 years old, with an average of 21.8 years old and a standard deviation of 2.15.
- The lack of male individuals is due to the Faculty where the experiment took place, since the percentage of male to female students is almost negligible.
- Therefore, no male candidates were included during collection of the database.
C. Task Justification
- Concretely, this paper proposes to induce stress by using hyperventilation (HV) and talk preparation (TP) [24].
- As a consequence, several physiological changes emerge, such as arterial-pressure diminution in blood until a certain level, so-called hypocapnea [24], [25], and blood pH increment, known as alkalosis.
- Voluntary HV does not always produce an actual anxiety reaction [24], and therefore, an additional anxiogenic task is required to ensure that a positive valence in terms of stress response is provoked.
- Results provided by Cano-Vindel et al. [24] and Zvolensky and Eifert [25] highlight that HV produces a physiological reaction (in terms of physiological registration, HR, and GSR) similar to the reaction induced by a threatening task of preparing a talk.
- These previous tasks have been widely studied and evaluated with positive results, and they are very suitable to induce stressing stimuli on individuals.
D. Procedure
- Two groups (namely, Group 1 and Group 2) were created, ensuring that the distribution of their respective anxiety levels, measured by psychological tests [26], [27], were similar.
- In other words, this selection seeks to avoid one group containing people which barely react against stress and another group with people which overreact under stressing conditions.
- Therefore, both groups must be well balanced in terms of anxiety levels in order to validate the experiments.
- Participants from Group 1 underwent an experimental session using physiological and subjective signals under the following conditions: calm state (baseline, namely, BL1), stimulating task (HV), threatening task (TP), and baseline poststress (BL2).
- The main reason for altering the order is to make the task order independent from the results obtained [24], [26].
E. Database Discussion
- Two questions arise from this database acquisition:.
- The former question has an affirmative answer.
- Moreover, stress mechanism extracts some information from the stimuli so that if such stressing agent appears again, the human body is able to react faster and better compared with the first time [1], [29].
- Finally, the latter question is hard to answer since it is difficult, even for expert psychologists, to state whether the response among female and male individuals differs, inasmuch as the previous response varies within female individuals [30].
V. STRESS-DETECTION SYSTEM
- A stress-detection system inherits several characteristics from biometric systems.
- This template is based on specific characteristics extracted from individual concerning parameters from the physiological signals HR and GSR.
- On the other hand, once the user is associated to a template, the individual is able to access the system, and therefore, a template comparison is required.
- Both steps are described in following sections.
A. Template Extraction
- Mathematically, both HR and GSR are considered as stochastic signals.
- Therefore, H represents the space of HR possible signals, and G represents the space of GSR possible signals.
- The decision to avoid normalization was done based on experience, since data without normalization provide more accurate results in terms of stress detection.
- This approach will facilitate the implementation of fuzzy antecedent membership functions by Gaussian distributions in a posterior fuzzy decision algorithm.
- Evidently, the performance of the system depends on this parameter since the longer tT is, the more information the system obtains, and therefore, the stress template may be more accurate.
B. Template Update
- These values do not remain unalterable but change along time.
- This template must be refreshed every time the user accesses the system so that the template can adapt to the variation of the specific individual.
- This paper proposes two possible system implementations: a manual implementation, where the system parameters are set without any learning procedure, and an automatic implementation, where the system adapts its internal parameters to relate input with output during training stage.
- In detail, the main difference between them relies on the fuzzy decision algorithm: First, manual implementation is based on a Mamdani [31] fuzzy decision algorithm using template T to describe antecedent membership functions and implementing the output with triangular distributions.
A. Manual Implementation
- As introduced previously, the manual implementation is designed with a Mamdani fuzzy system since it has been widely used in expert decision systems due to their comprehensible rules [31], [33].
- Furthermore, two triangular functions describing each output possibility (stress or nonstress) characterize the consequent membership functions.
- The triangular functions were selected according to their adequate properties to provide an accurate output [33].
- These rules are very intuitive, and they are provided by an expert, involving the previously described parameters ζhi , ζgi , σhi , and σgi .
- This state i represents the four situations corresponding to the experiments to induce stress: BL1, HV, TP, and BL2.
B. Automatic Implementation
- A different approach is used to improve the previous manual implementation since the algorithm learns which membership functions are to be used and which parameters must be selected in order to obtain a given output.
- This automatic implementation based on adaptive neuro-fuzzy inference system, which provides a Sugeno-type fuzzy inference system.
- In other words, Sugeno implementation considers as initial conditions the membership functions from the previous Mamdani implementation.
- This implementation needs a training stage to obtain the knowledge that is enough to properly detect stress [32].
C. Differences Between Implementations
- The main difference between the previous implementations (manual versus automatic) is related to the manner by which both systems are modeled.
- The Mamdani system models from the beginning both the output and the input and requires stresstemplate parameters (centroids and standard deviations).
- The output was modeled previously to the data, and that model was established manually based on a trial-and-error strategy.
- If the training stages require only data from BL1 and TP, then there will be only two rules, one corresponding to each stage.
- Finally, there is a difference in terms of performance, since the Sugeno implementation offers a higher success rate in differentiating stress from relax states.
D. Stress Measurement
- In other words, these systems can provide an output each second.
- The final value will result as a metric from this previous vector.
- These proposed measurements provide a result each tacq seconds, and, as discussed posteriorly in Section VII, measurements based on median μ1/2 usually produces a more accurate result when compared with μ measurements.
- This threshold, namely, ρth will be obtained so that the performance of the system is maximized.
A. Database: Training, Validation, and Testing Data
- In order to obtain valid results, the database must be divided into three groups.
- Used to fixed threshold ρth and temporal parameters (tT and tacq) in order to maximize the performance of the system.
- Notice that this validation scheme is similar to a K-fold cross validation.
- The justification for this division is based on the research carried out by Picard and Healey [18], where several physiological signals (not only HR and GSR) were recorded during a period of time of 32 days in the same person.
- Eight emotions were provoked during 30 min per day, and no substantial changes were appreciated during that period in each emotions.
B. Evaluation Schemes
- As a general idea and according to the procedure described in Section V-A, detecting stress involves the acquisition of physiological signals during four different stages: BL1, TP, HV, and BL2 (Section IV-D).
- In order to provide some light on this aspect, nine schemes are proposed with the sole purpose of answering the previous questions.
- Scheme considering both stressing and non- stressing stages, also known as 4) BL1 + TP.
- The subsequent experiments are based on this experimental setup.
C. True Stress Detection versus True Nonstress Detection
- A stress detection system must reach a compromise between detecting properly which individuals are under stress situations and which individuals are in a normal state of mind.
- This TSD factor corresponds to the sensitivity statistical measure since TSD can be described as follows: TSD = #True Positives #True Positives + #False Negatives (1) where a True Positive means classifying as stressed an individual which is indeed under stress, and False Negative means classifying as relaxed an individual which is under stressing situations.
- At this point, one question arises: TESD is obtained with validation data, and therefore, threshold ρth and temporal parameters tT and tacq are fixed to maximized TESD.
- D. Template Time (tT ) and Acquisition Time (tacq).
E. Evaluation Performance
- Once the previous parameters, i.e., threshold ρth, tT , and tacq are obtained in order to maximize TESD, the system is finally implemented.
- Fig. 2 shows visual information on the performance of the system under several schemes, highlighting the fact that the accuracy in detecting stress properly increases when BL2 is not considered in the training data.
- Furthermore, Fig. 2 shows the gathered TESD results corresponding to schemes BL1 + TP (98.9%, Sugeno μ1/2) and BL1 + HV (99.5%, Sugeno μ1/2), where only two stages were considered to train and validate the implementations.
- The results provided in Table I are obtained based on the average of the random experiments using a cross-validation approach.
- This is an outstanding result since it allows decreasing (in terms of time) the template-extraction step among the other aspects discussed in Section VIII.
Did you find this useful? Give us your feedback
Citations
260 citations
Cites background from "A Stress-Detection System Based on ..."
...To induce stress the authors in [14] propose using hyperventilation and talk preparation....
[...]
255 citations
194 citations
Cites methods from "A Stress-Detection System Based on ..."
...Thus, stress can be detected with the use of alvanic Skin response (GSR) which has been adopted as a relible psychophysical measure [25,26]....
[...]
186 citations
Cites background from "A Stress-Detection System Based on ..."
...[15] (2011) EDA, PPG Hyperventilation and Talk Prep Fuzzy Logic 2 (S, R) 99 Yes...
[...]
134 citations
Cites background from "A Stress-Detection System Based on ..."
...Evaluation of ANS activity can be performed by recording and analysing several physiological variables: heart rate [9], respiration rate (RespR) [10], ElectroEncephaloGraphical (EEG) activity [11], skin galvanic response (GSR) [12,13] and skin temperature [14]....
[...]
...The intended use for this ECG/EBI sensing unit is to record cardiogenic biopotentials to compute the HR from the acquired ECG and measure the impedance change caused during breathing to extract the RespR from the recorded changing TEB signal....
[...]
...Several textile-enabled measurement devices and sensorized garments have been implemented for recording physiological variables that would allow to study the response of the ANS during stressful tasks in a non-invasive manner e.g., parameters from cardiac activity, GSR dynamics, skin temperature and RespR [10,13,29,30]....
[...]
..., parameters from cardiac activity, GSR dynamics, skin temperature and RespR [10,13,29,30]....
[...]
...Nevertheless several wearable devices and sensorized garments have been implemented and used to record physiological variables to study the response of the ANS during stressful tasks in a non-invasive manner through, e.g., parameters from cardiac activity, GSR dynamics, skin temperature and RespR [10,13,29,30]....
[...]
References
5,868 citations
"A Stress-Detection System Based on ..." refers background in this paper
...individual, and such a response is similar, independent of the time during the stressing stimulus provoked the response [28]....
[...]
3,093 citations
"A Stress-Detection System Based on ..." refers methods in this paper
...fuzzy logic [6] and case-based reasoning [3]....
[...]
...Finally, the defuzzification method is carried out by a centroid method [6]....
[...]
1,777 citations
"A Stress-Detection System Based on ..." refers background or methods or result in this paper
...Notice that signals hi and gi are not normalized, in contrast to previous approaches [2], [8]....
[...]
...Moreover, the work presented by Healey and Picard [2] deserves special mention, since they are considered to be pioneers on stress-detection field....
[...]
...[18] W. Picard and J. A. Healey, “Wearable and automotive systems for affect recognition from physiology,” MIT, Cambridge, MA, Tech....
[...]
...[2] J. A. Healey and R. W. Picard, “Detecting stress during real-world driving tasks using physiological sensors,” IEEE Trans....
[...]
...The justification for this division is based on the research carried out by Picard and Healey [18], where several physiological signals (not only HR and GSR) were recorded during a period of time of 32 days in the same person....
[...]
953 citations
"A Stress-Detection System Based on ..." refers background in this paper
...Moreover, it is possible to deduce the intention of the individual from these results, a very remarkable conclusion for future computer applications and for the sake of a better human–computer interaction (HCI) [11], [12]....
[...]
652 citations
"A Stress-Detection System Based on ..." refers background in this paper
...In contrast to ECG biometric properties [21], HR is not distinctive enough to identify an individual....
[...]
Related Papers (5)
Frequently Asked Questions (7)
Q2. What have the authors stated for future works in "A stress-detection system based on physiological signals and fuzzy logic" ?
This template update will be considered in a future research. This implementation remains as future work. In other words, a wide variety of scenarios can benefit from this approach due to its noninvasiveness, the likelihood to be embedded on current security systems, and its possibility in detecting stress in real time [ 20 ], together with the capability of being combined to other stress-detection methods based on computer-vision algorithms. Moreover, future research entails integration with mobile devices.
Q3. What is the device used to carry out these experiments?
The device proposed to carry out these experiments is I-330-C2 PHYSIOLAB (J &J Engineering), which is able to process and store six channels, including electromyography, ECG, respiration rate, HR, and GSR.
Q4. What are the main characteristics of this system?
The main characteristics of this system is its noninvasiveness, fast-oriented implementation, and outstanding accuracy in detecting stress when compared with the previous approaches.
Q5. How many values were considered during this evaluation task?
In fact, only eight values were considered during this evaluation task for the sake of simplicity (tacq, tT ∈ {3, 5, 7, 10, 12, 15, 17, 20}) since TESD does not vary significatively between two values from the previous set and because more than 20 s is far beyond the sake of real applications.
Q6. What are the two temporal parameters that determine the performance of the system?
D. Template Time (tT ) and Acquisition Time (tacq)The performance of the system (TESD) not only depends on the previous threshold ρth but also on two temporal parameters: template time (tT ) and acquisition time (tacq).
Q7. What is the main characteristic of the work presented by Andren and Funk?
The work presented by Andren and Funk [3] provides a system that is able to compute the stress level of an individual by the manner and rhythm in which a person types characters on a keyboard or keypad.