Guan Xinping

Information-weighted Gaussian matrix in compressed sensing for ECG

An information-weighted Gaussian matrix (IWGM) is proposed to improve the quality of the recovered ECG signal from compressed sensing. In this paper, firstly we decrease the energy consumption by using the DCT transform instead of the traditional wavelet transform. Moreover, with the weighting of specific area in the sensing matrix, the IWGM decreases the overall error while the compressed ratio remain a relatively high level, thus, it guarantees a high-quality recovery. The experimental results indicate that the proposed matrix is suitable for compressed sensing of the ECG signal.

An overlapping distributed state estimation and detection method in smart grids

This paper proposes a novel distributed state estimation and detection algorithm in smart grids. By decomposing a whole power system into several overlapping interconnected areas, the centralized state estimation algorithm turns into a distributed state estimation algorithm. And by iteratively exchanging information with neighboring areas, the result of distributed state estimation can reach convergence and each subsystem can derive the states of the entire power system. When an attacker injects false data into measurements in any area, the neighboring honest areas can quickly detect this abnormality and decrease the mutual weights of their exchanging information between the suspicious area. When all the estimated state vectors converge, each control area can determine whether its neighboring area is intruded or not by using information from shared buses. The proposed approach not only proposes a distributed state estimation structure but also a detection method which has the capacity to detects false data injection (FDI) attacks. The performance of proposed algorithm is demonstrated on the IEEE 14-bus system.