Abstract—Many algorithms require prior information about test subjects for heart rate (HR) detection. In this paper, we propose a novel algorithm for detecting the HR of previously unknown species based on the Pan-Tompkins (PT) algorithm without a priori knowledge. In this improved PT algorithm, some parameters that need to be predefined can be adaptively adjusted according to the recognition of previous species. In the recognition step, a clustering algorithm is applied to obtain the rough RR intervals, and a decision tree is applied for species recognition by using two features: the rough RR intervals and the proportion of the ECG power of frequency that is less than 5 Hz. The accuracy and the Kappa of species classification in the recognition step can reach 93.15% and 90.23%, respectively. In the HR detection step, the improved PT algorithm is used to precisely detect the HR of rats, mice, humans, frogs and rabbits, and the results show that this method has good performance. In particular, we apply the proposed algorithm to test some ECG signals in humans from the MIT-BIH database. The results show that the accuracy of the proposed algorithm for detection of HR in humans reaches 99.71%.

Index Terms—Clustering algorithm, decision tree, pan-tompkins algorithm, RR-intervals, heart rate.

Dan Li, Tao Zhang, Zhonghui Jiang, and Defu Cheng are with Chengdu Techman Software Co., Ltd, Chengdu, China. Wu Huang is with Sichuan University, Chengdu, China. Guobiao Xu is with China Civil Aviation Flight University, Chengdu, China.

Cite:Dan Li, Wu Huang, Guobiao Xu, Tao Zhang, Zhonghui Jiang, Defu Cheng, "An Adaptive RR Interval Detection Algorithm Based on Species Recognition," International Journal of Computer and Electrical Engineering vol. 11, no. 4, pp. 180-191, 2019.