The paper presents the application of selected feature selection methods to determine their optimal set in the task of non-intrusive appliances identification. The process detects devices operating in the specific location based on the analysis of voltage and current signals. The informativeness calculation methods and Fisher linear discriminative analysis are proposed to find significant features for determining, which appliances changed their state. The approach is verified on two sets of features. The scheme was tested in the laboratory conditions with multiple appliances working simultaneously, exploiting the decision tree as the appliance identifier. Experimental results show usefulness of the approach maximizing the identification accuracy with decreased complexity of computations.

This paper deals with a method to limit the consequences of packet loss on the accuracy of a real-time human motion tracking system based on Wireless Sensor Network (WSN). Since the need for a prolonged battery life is a prominent commitment in a WSN design, the power budget is so low that communication turns out to be not reliable enough for real-time requirements. Whereas packets getting lost may not represent a big deal in most of WSN applications, they are very likely to cause troubles in a system supposed to sample, stream and compute under very tight constraints. For this reason, countermeasures have to be taken in order to tolerate packet loss and limit its impact on the body segments’ orientation estimation. The experimented method interpolates between quaternions produced by complementary algorithms merging smoothness of motion under ordinary circumstances with the capacity of fast recovery from loss. The results look promising and pave the way to further research for loss tolerance solution in low-power wireless systems.

In this paper we propose a new method for measurement of reactive power. The proposed method is based on stochastic digital measurement approach and discrete Fourier transform. The validity of the method is verified by numerous simulations. Besides this, it is shown that the proposed method is simple, robust and reliable, which makes it very suitable for practical implementation.

The article presents a new approach to identify bad power measurement based on the use of synchronized measurements of voltage and current phasors. The proposed method is based on the assumption that in most instances the bad power measurements are caused by inaccurate measurements of a current. In this case, the availability of voltage phasors’ measurements at consumers and in the main substation is a sufficient condition to identify bad power measurements. It has been proposed to use the instrument for state estimation theory and the recursive algorithm to identify good measurements of the currents’ and voltages’ phasors of the nodes in order to improve the reliability and accuracy of the method. The identification accuracy will depend primarily on the measurements synchronization accuracy and the passive parameters sizing error.

Autonomous underwater vehicles (AUV) perform inspection missions and intervention in known and unknown environments, where it is necessary to ensure their safety. The AUV must have the ability to detect and avoid obstacles in the path of navigation. This article, an obstacle detection system for experimental Guanay II AUV is proposed, using a mechanical scanning SONAR Tritech Micron MK3. Since Guanay II operates autonomously, we have designed software that allows adjustment and control of the parameters of SONAR, and the acquisition and processing of the signals obtained. Experimental tests at sea have allowed to verify the correct operation of software designed, as well as, experimental tests in a controlled environment have allowed to determine the optimal values of the basic parameters of SONAR.