Task 1A DCASE 2021: Acoustic Scene Classification with mismatch-devices using squeeze-excitation technique and low-complexity constraint

Acoustic scene classification (ASC) is one of the most popular problems in the field of machine listening. The objective of this problem is to classify an audio clip into one of the predefined scenes using only the audio data. This problem has considerably progressed over the years in the different editions of DCASE. It usually has several subtasks that allow to tackle this problem with different approaches. The subtask presented in this report corresponds to a ASC problem that is constrained by the complexity of the model as well as having audio recorded from different devices, known as mismatch devices (real and simulated). The work presented in this report follows the research line carried out by the team in previous years. Specifically, a system based on two steps is proposed: a two-dimensional representation of the audio using the Gamamtone filter bank and a convolutional neural network using squeeze-excitation techniques. The presented system outperforms the baseline by about 17 percentage points.

Design and analysis of transmission-level power system stability control

After the August 14, 2003 blackout, power system stability probelm has received a great deal of attention. This project is focused on the analysis of transient stability following disturbances. The project investigates three generator-excitation techniques for controlling the stability of power system. They are the manual control (constant excitation voltage), the automatic voltage regulator (AVR), and the automatic voltage regulator plus the power system stabilizer (PSS) that basically can be a led-leg compensator using a rotor speed deviation as an input feedback. This project carried out a computer simulation study of the power system stability for various disturbances conditions. The simulation results from the project have shown that AVR coupled PSS generator-excitation control can achieve the best power system stability, comparing to the manual or AVR control. A major contribution of this project is that a computer platform using MatLab/Simulink software was designed and can be used by other researchers for the power system stability study.

Design and analysis of transmission-level power system stability control

After the August 14, 2003 blackout, power system stability probelm has received a great deal of attention. This project is focused on the analysis of transient stability following disturbances. The project investigates three generator-excitation techniques for controlling the stability of power system. They are the manual control (constant excitation voltage), the automatic voltage regulator (AVR), and the automatic voltage regulator plus the power system stabilizer (PSS) that basically can be a led-leg compensator using a rotor speed deviation as an input feedback. This project carried out a computer simulation study of the power system stability for various disturbances conditions. The simulation results from the project have shown that AVR coupled PSS generator-excitation control can achieve the best power system stability, comparing to the manual or AVR control. A major contribution of this project is that a computer platform using MatLab/Simulink software was designed and can be used by other researchers for the power system stability study.

System Identification Guidance for Multirotor Aircraft: Dynamic Scaling and Test Techniques

State-space system identification was performed to extract flight dynamic models for hovering flight of a 55 cm, 1.56 kg hexacopter unmanned aerial vehicle. Different input excitation techniques were tested to determine which maneuvers provided high-quality system identification results for small-scale multirotor vehicles. These input excitation techniques included automated frequency sweeps, varying in amplitude, and multisine sweeps. Coherence, Cramer–Rao bounds, and insensitivities were used as metrics for comparing the system identification results. A parametric variation of frequency sweep amplitudes were performed in all axes (roll, yaw, pitch, and heave) to provide guidance on frequency sweep amplitude for small-scale multirotor unmanned aerial systems. The dynamics of the 55 cm hexacopter were used to estimate the dynamics of a larger 127-cm hexacopter via Froude scaling based on hub-to-hub distance as the characteristic length. The scaled results were compared to an actual system identification model of a 127-cm hexacopter.

Registration of high-frequency waves on the surface by the interference methods

Capillary waves are frequently used to measure the surface tension of liquids. However, this approach has not found wide application in the manufacture of modern commercial tensiometers because of the limitations imposed by capillary wave excitation techniques and the labor input associated with its practical implementation. In this paper we introduce a modified version of the capillary wave method which allows one to avoid the existing limitations and disadvantages. The distinguishing features of the proposed technique are as follows: acoustic wave generation and application of an interferometry technique for 3D surface profile reconstruction. A dynamic speaker with controlled vibration frequency and amplitude is used to produce acoustic vibrations. Application of a conventional Fizeau interferometer and the spatial phase shifting method makes it possible to perform surface form measurements with a high accuracy. For calculating wavelengths and the damping co-efficient, the surface profile is fitted with a decaying cylindrical wave equation. The accuracy of surface tension measurement by the modified capillary wave technique is 0.3 %. Owing to the non-contact way of wave generation and the small amounts of the examined fluid, the proposed method can be used in different studies.

Ionization dynamics of a phenylenediamine derivative in solutions as revealed by femtosecond simultaneous and stepwise two-photon excitation

Photoionization dynamics of TMPD in solution has been investigated by means of multi-photon excitation techniques.