CHAR-HMM: An Improved Continuous Human Activity Recognition Algorithm Based on Hidden Markov Model
With the rapid development of wearable sensor technology, Human Activity Recognition (HAR) based on sensor data has attracted more and more attentions. The Hidden Markov Model (HMM) with perfect performance in speech recognition has a good effect on HAR. However, almost all these techniques train multiple Hidden Markov Models for different classes of activity. For a given activity sequence with multiple activities, the activity corresponding to the HMM with the maximum generating probability is selected as the recognition result, which is not suitable for continuous HAR with multiple activities. For this problem, we propose an improved Hidden Markov activity recognition algorithm where discriminative model and generative model are utilized. The discriminative model SVM is used to produce the observation sequence of HMM, and the generative model HMM is used to generate the final result. Compared with the traditional Hidden Markov HAR model, our proposal has good performance in terms of precision, recall and F1 score.
KeywordsHuman Activity Recognition Hidden Markov Model Support Vector Machine Cyber-Physical system Accelerometer signal
The corresponding author Botao Wang is supported by the NSFC (Grant No. 61173030).
- 7.Beily, M.D.E., Badjowawo, M.D., Bekak, D.O., et al.: A sensor based on recognition activities using smartphone. In: 2016 International Seminar on Intelligent Technology and Its Applications (ISITIA), pp. 393–398. IEEE (2016)Google Scholar
- 8.Kurban, O.C., Yildirim, T.: Neural network based daily activity recognition without feature extraction. In: 2014 22nd Signal Processing and Communications Applications Conference (SIU), pp. 567–570. IEEE (2014)Google Scholar
- 9.Margarito, J., Helaoui, R., Bianchi, A.M., et al.: User-independent recognition of sports activities from a single wrist-worn accelerometer: a template-matching-based approach. IEEE Trans. Biomed. Eng. 63(4), 788–796 (2016)Google Scholar
- 10.Piyathilaka, L., Kodagoda, S.: Gaussian mixture based HMM for human daily activity recognition using 3D skeleton features. In: 2013 8th IEEE Conference on Industrial Electronics and Applications (ICIEA), pp. 567–572. IEEE (2013)Google Scholar
- 11.Nickel, C., Busch, C., Rangarajan, S., et al.: Using hidden markov models for accelerometer-based biometric gait recognition. In: 2011 IEEE 7th International Colloquium on Signal Processing and its Applications (CSPA), pp. 58–63. IEEE (2011)Google Scholar
- 12.Cheng, L., Guan, Y., Zhu, K., et al.: Recognition of human activities using machine learning methods with wearable sensors. In: 2017 IEEE 7th Annual Computing and Communication Workshop and Conference (CCWC), pp. 1–7. IEEE (2017)Google Scholar
- 13.Nickel, C., Brandt, H., Busch, C.: Benchmarking the performance of SVMs and HMMs for accelerometer-based biometric gait recognition. In: 2011 IEEE International Symposium on Signal Processing and Information Technology (ISSPIT), pp. 281–286. IEEE (2011)Google Scholar
- 14.Wang, J., Chen, R., Sun, X., et al.: Generative models for automatic recognition of human daily activities from a single triaxial accelerometer. In: The 2012 International Joint Conference on Neural Networks (IJCNN), pp. 1–6. IEEE (2012)Google Scholar
- 15.Quinlan, J.R.: Induction of decision trees. Mach. Learn. 1(1), 81–106 (1986)Google Scholar
- 17.Stauffer, C., Grimson, W.E.L.: Adaptive background mixture models for real-time tracking. In: IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2, p. 252. IEEE Xplore (1999)Google Scholar