Antonio Fasano

Baseline wander removal for bioelectrical signals by quadratic variation reduction

Baseline wander is a low-frequency additive noise affecting almost all bioelectrical signals, in particular the ECG. In this paper, we propose a novel approach to baseline wander estimation and removal for bioelectrical signals, based on the notion of quadratic variation reduction. The quadratic variation is meant as a measure of variability for vectors or sampled functions, and is a consistent measure in this regard. Baseline wander is estimated solving a constrained convex optimization problem where quadratic variation enters as a constraint. The solution depends on a single parameter whose value is not critical, as proven by a sensitivity analysis. Numerical results confirm the effectiveness of the approach, which outperforms state-of-the-art algorithms. The algorithm compares favorably also in terms of computational complexity, which is linear in the size of the vector to detrend. This makes it suitable for real-time applications as well as for applications on devices with reduced computing power, such as handheld devices.

Baseline wander estimation and removal by quadratic variation reduction

The baseline wander is a low frequency additive noise partially overlapping the band of ECG signal. This makes its removal difficult without affecting the ECG. In this work we propose a novel approach to baseline wander estimation and removal based on the notion of quadratic variation. The quadratic variation is a suitable index of variability for vectors and sampled functions. We derive an algorithm for baseline estimation solving a constrained convex optimization problem. The computational complexity of the algorithm is linear in the size of the ECG record to detrend, making it suitable for realtime applications. Simulation results confirm the effectiveness of the approach and highlight its ability to remove baseline wander. Eventually, the proposed algorithm is not limited to ECG signals, but can be effectively applied whenever baseline estimation and removal are needed, such as EEG records.

Reduced-order quadratic Kalman-like filtering for non-Gaussian systems

In this paper the state estimation problem for linear discrete-time systems with non-Gaussian state and output noises is treated. In order to obtain a state optimal quadratic estimate with a lower computational effort and without loosing the stability, only the observable part of the second-order power system will be considered. The novelty of the proposed algorithm is to provide a method to compute, in a closed form, the rank of the observability matrix for the quadratic system. Considering a new augmented state-space built as the aggregate of the actual state vector and the observable components of the system squared state, and defining a new observation sequence composed of the original output measurements together with their square values, we will be in a condition to use Kalman filtering that, in this case, produces a suboptimal quadratic stable state estimate for the original system. The solution is given in closed form by a recursive algorithm.

Fast ECG baseline wander removal preserving the ST segment

Baseline wander removal is an unavoidable preprocessing step in ECG signal analysis. Unfortunately, the in-band nature of this kind of noise makes its removal difficult without affecting ECG waveform, in particular the ST segment. This is a portion of the ECG with high clinical relevance, as it is related to the diagnosis of acute coronary syndromes. The ST segment is highly susceptible to distortion when baseline removal is performed affecting the low-frequency region of ECG spectrum, where are concentrated the harmonic components that mainly contribute to the shape of the ST segment. In this paper, we propose to tackle the problem of baseline removal from a different perspective, considering the quadratic variation as an alternative measure of variability not directly related to the frequency domain. In this regard, we recently proposed a novel baseline removal algorithm based on quadratic variation reduction. In this paper, we assess its performance with respect to the distortion of the ST segment comparing it to state-of-the-art algorithms. Simulation results confirm the effectiveness of the approach based on quadratic variation reduction. Our algorithm outperforms state-of-the-art algorithms tailored to minimize distortion of the ST segment. Moreover, it compares favorably also in terms of computational complexity, which is linear in the size of the vector to detrend. This makes it suitable also for real-time applications.

Denoising and harmonic artifacts rejection for ECG P-waves by quadratic variation reduction

Atrial fibrillation (AF) is a common cardiac arrhythmia related to irregular atrial contractions. Several studies have shown that the analysis of P-waves extracted from ECG signals is helpful in understanding the predisposing factors to AF. However, P-waves are usually highly corrupted by noise and harmonic artifacts and this makes quite difficult their analysis. Recently we proposed a novel algorithm for denoising P-waves based on the notion of quadratic variation reduction. It is quite good in denoising P-waves affected by noise, but its effectiveness reduces when it is used in filtering out harmonic artifacts, like power-line interference. In this paper we propose an algorithm that overcomes this limitation and extends our previous method allowing it to both denoise and reject harmonic artifacts. Simulation results confirm the effectiveness of the approach and highlight its ability to remove both noise and artifacts. The algorithm has reduced computational complexity and this makes it suitable for real-time applications.

A detection-estimation approach to filtering for Gaussian systems with intermittent observations

In this paper we consider the problem of state estimation for linear discrete-time Gaussian systems with intermittent observations. Intermittent observations result from packet dropouts when data travel along unreliable communication channels, as in the case of wireless sensor networks, or networked control systems. We assume that the receiver does not know the sequence of packet dropouts, which is a common situation, e.g., in wireless sensor networks or in networks that cannot rely on protocols that provide information on packet loss. In this paper we propose a detection-estimation approach to the problem of state estimation. The estimator consists of two stages: the first is a nonlinear optimal detector, which decides if a packet dropout has occurred, and the second is a time-varying Kalman filter, which is fed with both the observations and the decisions from the first stage. The overall estimator has finite memory and the tradeoff between performance and computational complexity can be easily controlled. Simulation results highlight the effectiveness of the proposed approach, which outperforms the linear optimal filter of Nahi. Finally, the method is amenable to generalization.

ECG baseline wander removal by QVR preserving the ST segment

Baseline wander removal is an unavoidable step in ECG signal processing. The in-band nature of this noise makes its removal difficult without affecting the ECG, in particular the ST segment. This portion of the ECG has high clinical relevance, as it is related to the diagnosis of acute coronary syndromes. We have recently proposed a novel approach to baseline wander removal based on the notion of quadratic variation reduction. In this paper, we assess its performance in terms of both effectiveness in removing baseline wander and distortion introduced in the ST segment. Numerical results highlight the effectiveness of the approach, which outperforms state-of-the-art algorithms both in removing baseline drift and preserving the ST segment. The algorithm is also very fast, as its computational complexity is linear in the size of the vector to detrend.

ECG smoothing and denoising by local quadratic variation reduction

The ECG is the standard noninvasive test used to measure the electrical activity of the heart. Unfortunately, ECG signal is corrupted by several kinds of noise and artifacts that may negatively affect any subsequent analysis. In this work, we present a fast and effective algorithm for smoothing and denoising ECG records. The algorithm is the closed-form solution to a constrained convex optimization problem, where smoothing and denoising are achieved by locally reducing the quadratic variation of different portions of the ECG. Such a reduction is inversely related to the local SNR. The computational complexity of the algorithm is linear in the size of the vector under analysis, thus making it suitable for real-time applications. Simulation results confirm the effectiveness of the approach and highlight a notable ability to smooth and denoise ECG signals.

A thruster failure tolerant control scheme for underwater vehicles

This paper extends our previous work [1] on the design of a thruster failure tolerant control scheme for underwater vehicles. The proposed control scheme is based on the use of a suitable thruster allocation algorithm, which consists on a modified version of the Moore-Penrose pseudo inverse. In this work, each thruster of the underwater vehicle can rotate, offering a significant advantage to optimize its control. When a thruster experiences a failure, the resulting thrust force, which should be allocated to the failed actuator, is reallocated to the still faultless thrusters. Moreover, the angle of each thrusters is set to minimize the control effort. A bank of controllers is built so that each controller is designed to control the considered underwater vehicle under a specific actuator failure scenario.

A detection-estimation approach to filtering with intermittent observations with generally correlated packet dropouts

This paper is concerned with the problem of state estimation for the class of linear discrete-time Gaussian systems with intermittent observations due to packet losses. This is a common case in networked control systems, where the state of a remote plant is estimated from measurements carried through a lossy network. We assume that the receiver does not know the sequence of packet dropouts. This is typical, e.g., in wireless sensor networks or in networks that cannot rely on protocols that provide information on packet loss. Moreover, we assume that the sequence of packet dropouts is correlated, thus subsuming both the cases of independent dropouts and dropouts modeled as a Markov chain. We propose a detection-estimation approach to the problem of state estimation. The estimator consists of two stages: the first is a nonlinear optimal detector, which decides if a packet dropout has occurred, and the second is a time-varying Kalman filter, which is fed with both the observations and the decisions from the first stage. The overall estimator has finite memory and the tradeoff between performance and computational complexity can be easily controlled. As a case study, we derive the decision rule in closed form in the case of dropout sequence modeled as a Markov chain. Simulation results highlight the effectiveness of the proposed approach, which outperforms the linear recursive estimator of Hadidi and Schwartz.