A COLLECTIVE BIOLOGICAL PROCESSING ALGORITHM FOR ECG SIGNALS Horia Mihail Teodorescu FAS, Harvard University, Cambridge, MA, U.S.A. Keywords: Swarm model, Signal processing, Filtering, Biologic signal, EKG. Abstract: We establish and explore an analogy between hunting by packs of agents and signal processing. We present a version of adaptive ‘Hunting Swarm’ algorithm (HSA), apply it to ECG signals, and investigate the influence of the model parameters on the filtering of stationary and nonstationary noise. We show that results obtained with the HSA filter may outperform results obtained with several other filters. 1 INTRODUCTION Biological signals have wide bandwidth and may be affected by various noises. The first stage in process- ing such signals consists of filtering them in order to achieve a good signal to noise ratio (SNR). This task is often challenging because of the wide band of the signals and of the noises. As a consequence, nu- merous papers have been published recently propos- ing new filtering methods for ECG signals (Almenar and Albiol, 1999), (Leski and Henzel, 2005), (Ko- tas, 2007), (Korrek and Nizam, 2010), (Bansal et al., 2009), (Yan et al., 2010). In a previous communication (Teodorescu and Malan, 2010), we introduced an image processing al- gorithm based on swarms. In this paper we explore several variants of the ‘hunting swarm algorithm’ (HSA) and analyze their ability to remove noise from EKG signals for various signal to noise ratios (SNR). The signal is ‘enacted’ by the trajectory of a prey hunted by the swarm, as detailed in section 2. The models of the swarms in this paper include salient features from various swarm models reported in the literature and features that we introduced based on general considerations or from experimentation with model parameters. The organization of the paper is as follows. In the second section we expose the method to transform the signal processing task into a pack-hunting-a-prey task and describe the equations describing the prey and the pack movements. The third section is devoted to the results of filtering ECG signals with the HSA algo- rithm. The details of the implementation and the re- sults are discussed in the fourth section. Conclusions are drawn in the last section. 2 THE HS SIGNAL PROCESSING METHOD 2.1 Metaphor of the Hunting Pack In this section we suggest and exploit an analogy be- tween signal filtering and the natural hunting packs. We use this analogy to produce an algorithm for non- linear signal processing. The analogy has two main players: the prey and the hunting pack. The prey does not collaborate to the signal processing; instead, it en- acts the signal. The pack performs a virtual hunting and in so doing it produces the output (processed) sig- nal as the trajectory of the center of the pack. The hunting pack model, while borrowing much from var- ious swarm models, has many new features that give reason to consider it a new swarm model. Figure 1 depicts a sketch of a simplified process- ing procedure. In this sketch, the swarm is assumed constrained on a line at each time moment, with the agents taking positions along that vertical line, ac- cording to movement equations governed by inter- agent forces and to agent to prey forces. The prey moves in discrete time along the signal. The agents are attracted by the prey, thus tending to follow the prey. Consequently, the center of the swarm describes a trajectory in the plane. That trajectory is the result of ‘processing’ the prey trajectory, i.e. the signal, by the swarm. 413 Mihail Teodorescu H.. A COLLECTIVE BIOLOGICAL PROCESSING ALGORITHM FOR ECG SIGNALS. DOI: 10.5220/0003136304130420 In Proceedings of the International Conference on Bio-inspired Systems and Signal Processing (BIOSIGNALS-2011), pages 413-420 ISBN: 978-989-8425-35-5 Copyright c  2011 SCITEPRESS (Science and Technology Publications, Lda.)