Panic Detection Using Machine Learning and Real-Time Biometric and Spatiotemporal Data
Abstract
:1. Introduction
2. Panic Behavior and Sensing
2.1. Panic
2.2. Mobile Crowdsensing
3. Methodology
3.1. Dataset
- 1.
- Biometric (from wearable): heart rate, heart rate variability
- 2.
- Spatiotemporal (from smartphone): location coordinates, activity, speed, steps
- 3.
- Descriptive (from wearable): gender, age, weight
- 4.
- Secure ID (from smartphone)
3.2. Dataset Scenarios
4. Experimental Setup and Results
4.1. Machine Learning Classifiers
4.2. Deep Learning
4.3. Discussion
5. Conclusions
Author Contributions
Funding
Data Availability Statement
Conflicts of Interest
References
- Rose, D.A.; Murthy, S.; Brooks, J.; Bryant, J. The evolution of public health emergency management as a field of practice. Am. J. Public Health 2017, 107, S126–S133. [Google Scholar] [CrossRef] [PubMed]
- Aguirre, B.E. Emergency evacuations, panic, and social psychology. Psychiatry 2005, 68, 121–129. [Google Scholar] [CrossRef] [PubMed]
- Fact Sheets on the European Union—2019. 1. Available online: www.europarl.europa.eu/factsheets/en (accessed on 1 August 2022).
- Bui, T.; Sankaran, S. Foundations for Designing Global Emergency Response Systems (ERS). In Proceedings of the 3rd International ISCRAM Conference, Newark, NJ, USA, 13–17 May 2006; pp. 72–81. [Google Scholar]
- Li, N.; Sun, M.; Bi, Z.; Su, Z.; Wang, C. A new methodology to support group decision-making for IoT-based emergency response systems. Inf. Syst. Front. 2014, 16, 953–977. [Google Scholar] [CrossRef]
- Helbing, D.; Farkas, I.; Vicsek, T. Simulating dynamical features of escape panic. Nature 2000, 407, 487–490. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Hao, Y.; Xu, Z.; Wang, J.; Liu, Y.; Fan, J. An Approach to Detect Crowd Panic Behavior using Flow based Feature. In Proceedings of the 22nd International Conference on Automation and Computing, Colchester, UK, 7–8 September 2016; ISBN 9781862181328. [Google Scholar] [CrossRef] [Green Version]
- Ammar, H.; Cherif, A. DeepROD: A deep learning approach for real-time and online detection of a panic behavior in human crowds. Mach. Vis. Appl. 2021, 32, 57. [Google Scholar] [CrossRef]
- Yu, Y.; Si, X.; Hu, C.; Zhang, J. A Review of Recurrent Neural Networks: LSTM Cells and Network Architectures. Neural Comput. 2019, 31, 1235–1270. [Google Scholar] [CrossRef]
- Sufri, S.; Dwirahmadi, F.; Phung, D.; Rutherford, S. A systematic review of Community Engagement (CE) in Disaster Early Warning Systems (EWSs). Prog. Disaster Sci. 2019, 5, 100058. [Google Scholar] [CrossRef]
- United Nations for Disaster Risk, Reduction Disaster Preparedness for Effective Response, Guidance and Indicator Package for Implementing Priority Five of the Hyogo Framework: Hyogo Framework for Action, Building the Resilience of Nations and Communities to Disasters; UN Publication: Geneva, Switzerland, 2008; pp. 2005–2015. [CrossRef]
- United Nations. Global Survey of Early Warning Systems: An Assessment of Capacities, Gaps and Opportunities towards Building a Comprehensive Global Early Warning System for All Natural Hazards; A Global Report; United Nations (UN): Bonn, Germany, 2006. [Google Scholar]
- Basher, R. Global early warning systems for natural hazards: Systematic and people-centred. Philos. Trans. R. Soc. Lond. Ser. A Math. Phys. Eng. Sci. 2006, 364, 2167–2182. [Google Scholar] [CrossRef]
- United Nations for Disaster Risk Reduction. Sendai Framework for Disaster Risk Reduction 2015–2030; UN Publication: Geneva, Switzerland, 2015. [Google Scholar] [CrossRef]
- Liu, J.; Chen, Y.; Chen, Y. Emergency and disaster management-crowd evacuation research. J. Ind. Inf. Integr. 2021, 21, 100191. [Google Scholar] [CrossRef]
- Li, L. Education supply chain in the era of Industry 4.0. Syst. Res. Behav. Sci. 2020, 37, 579–592. [Google Scholar] [CrossRef]
- Lin, Y.; Duan, X.; Zhao, C.; Xu, L. Systems Science Methodological Approaches; CRC Press: Boca Raton, FL, USA; Taylor & Francis: Abingdon, UK, 2012; ISBN 978-1-4398-9551-1. [Google Scholar]
- Xu, L. Introduction: Systems science in industrial sectors. Syst. Res. Behav. Sci. 2013, 30, 211–213. [Google Scholar] [CrossRef]
- Xu, L.D. The contribution of systems science to Industry 4.0. Syst. Res. Behav. Sci. 2020, 37, 618–631. [Google Scholar] [CrossRef]
- Xu, L.; Cai, L.; Zhao, S.; Ge, B. Editorial: Inaugural Issue. J. Ind. Integr. Manag. 2016, 1, 1601001. [Google Scholar] [CrossRef] [Green Version]
- Tsai, C.-H.; Chen, P.-C.; Liu, D.-S.; Kuo, Y.-Y.; Hsieh, T.-T.; Chiang, D.-L.; Lai, F.; Wu, C.-T. Panic attack prediction using wearable devices and machine learning: Development and cohort study. JMIR Med. Inform. 2022, 10, e33063. [Google Scholar] [CrossRef] [PubMed]
- Chair-Mase, K.G.; Chair-Langheinrich, M.G.; Chair-Gatica-Perez, D.G.; Chair-Gellersen, H.P.; Chair-Choudhury, T.P.; Chair-Yatani, K.P. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing. Osaka, Japan, 7–11 September 2015. [CrossRef]
- Kutsarova, V.; Matskin, M. Combining Mobile Crowdsensing and Wearable Devices for Managing Alarming Situations. In Proceedings of the 2021 IEEE 45th Annual Computers, Software, and Applications Conference (COMPSAC), Madrid, Spain, 12–16 July 2021; pp. 538–543. [Google Scholar] [CrossRef]
- Alsalat, G.Y.; El-Ramly, M.; Aly, A.; Said, K. Detection of Mass Panic using Internet of Things and Machine Learning. Int. J. Adv. Comput. Sci. Appl. 2018, 9, 320–329. [Google Scholar] [CrossRef]
- Available online: https://dictionary.cambridge.org/dictionary/english/panic (accessed on 1 August 2022).
- Lin, P.; Ma, J.; Lo, S. Discrete element crowd model for pedestrian evacuation through an exit. Chin. Phys. B 2016, 25, 034501. [Google Scholar] [CrossRef] [Green Version]
- Bracha, H.S. Freeze, flight, fight, fright, faint: Adaptationist perspectives on the acute stress response spectrum. CNS Spectr. 2004, 9, 679–685. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Haag, A.; Goronzy, S.; Schaich, P.; Williams, J. Emotion recognition using bio-sensors: First steps towards an automatic system. In Tutorial and Research Workshop on Affective Dialogue Systems; Springer: Berlin/Heidelberg, Germany, 2004; pp. 36–48. [Google Scholar] [CrossRef]
- Kandel, E.R.; Schwartz, J.H.; Jessell, T.J. (Eds.) Principles of Neural Science, 4th ed.; McGraw-Hill: New York, NY, USA, 2000. [Google Scholar]
- Lang, P.J.; Bradley, M.M.; Cuthbert, B.N. International Affective Picture System (IAPS): Instruction Manual and Affective Ratings; Technical Report A-5; The Center for Research in Psychophysiology, University of Florida: Gainesville, FL, USA, 2001. [Google Scholar]
- Ganti, R.K.; Ye, F.; Lei, H. Mobile crowdsensing: Current state and future challenges. IEEE Commun. Mag. 2011, 49, 32–39. [Google Scholar] [CrossRef]
- Mehta, U.; Soni, P.; Fiaidhi, J. MobileCrowdSensing (MCS). TechRxiv Powered By IEEE. 9 April 2020. Available online: https://doi.org/10.36227/techrxiv.12101277 (accessed on 3 October 2022).
- Saquib, N.; Papon, T.I.; Ahmad, I.; Rahman, A. Measurement of heart rate using photoplethysmography. In Proceedings of the 2015 International Conference on Networking Systems and Security (NSysS), Dhaka, Bangladesh, 5–7 January 2015; pp. 1–6. [Google Scholar] [CrossRef]
- Nath, R.K.; Thapliyal, H.; Caban-Holt, A.; Mohanty, S.P. Machine Learning Based Solutions for Real-Time Stress Monitoring. IEEE Consum. Electron. Mag. 2020, 9, 34–41. [Google Scholar] [CrossRef]
- Allen, J. Photoplethysmography and its application in clinical physiological measurement. Physiol. Meas. 2007, 28, R1–R39. [Google Scholar] [CrossRef] [Green Version]
- Available online: https://www.forbes.com/health/healthy-aging/normal-heart-rate-by-age/ (accessed on 10 August 2022).
- Centers for Disease Control Website. Target Heart Rate and Estimated Maximum Heart Rate. Available online: https://www.cdc.gov/physicalactivity/basics/measuring/heartrate.htm (accessed on 10 August 2022).
- Loh, W.-Y. Fifty Years of Classification and Regression Trees. Int. Stat. Rev. 2014, 82, 329–348. [Google Scholar] [CrossRef] [Green Version]
- Hosmer, D.W., Jr.; Lemeshow, S.; Sturdivant, R.X. Applied Logistic Regression; John Wiley & Sons: Hoboken, NJ, USA, 2013; Volume 398. [Google Scholar] [CrossRef]
- Ren, J.; Lee, S.D.; Chen, X.; Kao, B.; Cheng, R.; Cheung, D. Naive bayes classification of uncertain data. In Proceedings of the 2009 Ninth IEEE International Conference on Data Mining, Miami Beach, FL, USA, 6–9 December 2009; pp. 944–949. [Google Scholar] [CrossRef]
- Keerthi, S.S.; Lin, C.-J. Asymptotic Behaviors of Support Vector Machines with Gaussian Kernel. Neural Comput. 2003, 15, 1667–1689. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Elith, J.; Leathwick, J.R.; Hastie, T. A working guide to boosted regression trees. J. Anim. Ecol. 2008, 77, 802–813. [Google Scholar] [CrossRef] [PubMed]
- Chandrashekar, G.; Sahin, F. A survey on feature selection methods. Comput. Electr. Eng. 2014, 40, 16–28. [Google Scholar] [CrossRef]
Feature | Type | Values | Units |
---|---|---|---|
Heart Rate | Numeric | 56 to 246 | bpm |
Heart Rate Variability | Numeric | 244 to 1071 | msec |
Activity | Categorical | Still, on foot, running, in vehicle | - |
Speed | Numeric | 0 to 58.3 | km/h |
Steps | Numeric | 0 to 240 | steps/min |
Gender | Categorical | male, female | - |
Age | Numeric | 20 to 84 | years |
Weight | Numeric | 50 to 90 | kg |
HRMAD10 | Numeric | −24 to 40 | bpm |
HRMAD30 | Numeric | −42 to 58 | bpm |
HRMAD60 | Numeric | −55 to 86 | bpm |
Scenario | Description | Example |
---|---|---|
Panic Scenario 1 | While being still, the subject starts running suddenly in panic as if an escape attempt takes place. | The subject rests when suddenly an alarming situation occurs. It stands up immediately and starts running out in the street. |
Panic Scenario 2 | While being on foot, the subject starts running suddenly in panic as if an escape attempt takes place. | The subject walks in the street and suddenly starts running because an attack takes place nearby. |
Panic Scenario 3 | While running for exercise the subject starts running in panic and its biometric characteristics start to deviate. | The subject runs at night for exercise when suddenly a wild animal starts chasing him. The subject begins to accelerate in an attempt to escape. |
Panic Scenario 4 | While being still, the subject’s biometric characteristics start to deviate but the subject freezes to its location. | The subject is still, and an unpleasant event happens. It panics and freezes in its location. When it calms down, it walks away from the scene. |
Panic Scenario 5 | While being on foot, the subject’s biometric characteristics start to deviate but it keeps walking as if it is not able to run. | The subject is walking when an attack takes place nearby. It panics but it remains in a walking state because it is slightly wounded. |
Panic Scenario 6 | While being in a vehicle, the subject’s biometric characteristics start to deviate, it stops the vehicle and starts running. | The subject is driving when an explosion happens nearby. All the vehicles are immobilized so it leaves the vehicle and escapes the scene running. |
Calm Scenario 1 | While being still, the subject suddenly starts running. | The subject is resting but suddenly remembers to do something, so it starts running. |
Calm Scenario 2 | While being on foot, the subject’s biometric characteristics start to deviate but it keeps walking. | The subject is walking on the street but suddenly develops palpitation due to cardiac causes. |
Calm Scenario 3 | While driving, the subject stops the vehicle and starts running. | While the subject is driving a car, it parks it and starts running to catch the bus. |
Classifier Model | Hyperparameters |
---|---|
Decision Tree | Max number of splits: 100, Split criterion: Gini’s diversity index, Surrogate decision splits: Off, Max surrogates per node: 10 |
Logistic Regression | No hyperparameter options |
Gaussian Naïve Bayes | Numeric predictors distribution: Gaussian, Categorical predictors distribution: MVMN |
Kernel Naïve Bayes | Numeric predictors distribution: Kernel, Categorical predictors distribution: MVMN, Kernel Type: Gaussian, Support: Unbounded |
Gaussian SVM | Kernel Function: Gaussian, Kernel scale: 0.61, Box constraint level: 1, Multiclass Method: One-vs-One, Standardize data: True |
SVM Kernel | Learner: SVM, Number of expansion dimensions: Auto, Lambda: Auto, Kernel scale: Auto, Multiclass Method: One-vs-One, Iteration limit: 1000 |
Boosted Trees | Ensemble method: Adaboost, Learner type: Decision Tree, Max number of splits: 20, Number of learners: 30, Learning rate: 0.1, Number of predictors to sample: All |
Classification Model | Accuracy (Raw Features) |
---|---|
Decision Tree | 91.4% |
Logistic Regression | 89.6% |
Gaussian Naïve Bayes | 79.6% |
Kernel Naïve Bayes | 80.3% |
Gaussian SVM | 93.2% |
SVM Kernel | 90.6% |
Boosted Trees | 90.9% |
Classification Model | Accuracy (Raw Features + HRMAD10) | Accuracy (Raw Features + HRMAD30) | Accuracy (Raw Features + HRMAD60) |
---|---|---|---|
Decision Tree | 93.3% | 92.8% | 92.8% |
Logistic Regression | 89.5% | 89.0% | 89.5% |
Gaussian Naïve Bayes | 80.4% | 80.6% | 81.3% |
Kernel Naïve Bayes | 83.7% | 84.3% | 85.3% |
Gaussian SVM | 94.2% | 94.4% | 94.5% |
SVM Kernel | 91.8% | 92.3% | 94.1% |
Boosted Trees | 92.3% | 93.3% | 93.9% |
Feature | MRMR | χ2 | ANOVA |
---|---|---|---|
HRV | 1 | 2 | 1 |
HRMAD60 | 2 | 1 | 2 |
Heart Rate | 4 | 3 | 3 |
Activity | 3 | 5 | 6 |
Speed | 6 | 4 | 5 |
Steps | 5 | 6 | 4 |
Classifier Model | Accuracy (HRV) | Accuracy (HRV + HRMAD60) |
---|---|---|
Decision Tree | 87.0% | 91.5% |
Logistic Regression | 86.1% | 88.2% |
Gaussian Naïve Bayes | 86.6% | 87.3% |
Kernel Naïve Bayes | 86.3% | 89.1% |
Gaussian SVM | 87.0% | 90.7% |
SVM Kernel | 78.0% | 79.7% |
Boosted Trees | 87.3% | 90.7% |
Approach | Accuracy |
---|---|
Raw Features Only | 91.5% |
Raw Features + HRMAD10 | 91.8% |
Raw Features + HRMAD30 | 92.3% |
Raw Features + HRMAD60 | 93.4% |
Sequence Length | Accuracy |
---|---|
1 sample | 90.6% |
10 samples | 90.8% |
30 samples | 91.4% |
60 samples | 94.0% |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Lazarou, I.; Kesidis, A.L.; Hloupis, G.; Tsatsaris, A. Panic Detection Using Machine Learning and Real-Time Biometric and Spatiotemporal Data. ISPRS Int. J. Geo-Inf. 2022, 11, 552. https://doi.org/10.3390/ijgi11110552
Lazarou I, Kesidis AL, Hloupis G, Tsatsaris A. Panic Detection Using Machine Learning and Real-Time Biometric and Spatiotemporal Data. ISPRS International Journal of Geo-Information. 2022; 11(11):552. https://doi.org/10.3390/ijgi11110552
Chicago/Turabian StyleLazarou, Ilias, Anastasios L. Kesidis, George Hloupis, and Andreas Tsatsaris. 2022. "Panic Detection Using Machine Learning and Real-Time Biometric and Spatiotemporal Data" ISPRS International Journal of Geo-Information 11, no. 11: 552. https://doi.org/10.3390/ijgi11110552