Gheorghe Serban

A vector approach to transient signal processing

The detection and characterization of burst signals are challenging tasks for time-frequency analysis, due to their very short duration. This paper investigates in this context the recurrence plot analysis (RPA) method, from which it derives the vector samples processing (VeSP) concept. The paper shows that VeSP is a generic framework that unifies signal processing concepts like histogram and autocorrelation, which it also generalizes and extends. Results of VeSP based tools are provided, concerning detection of transient signals, noise reduction, and frequency estimation.

On the recurrence plot analysis method behaviour under scaling transform

In the last decade, the applications of the recurrence plot analysis method make it a valuable alternative to the time-frequency and time-scale tools. As it was initially developed for the study of dynamical systems, and was later used in nonlinear time series analysis, the question of using it as a signal processing tool has not been put into discussion yet. In this field the projective techniques are largely used, with good results. Nevertheless, they also have some limitations — especially regarding transient signal processing. But this kind of signals are ubiquitous in real world. In addition, propagation through various media as well as on multiple paths lead to delayed, attenuated and dilated versions of the original transients. In this paper we study the behaviour of the recurrence plot analysis method in the context of analyzing some finite duration signals being subject to rescalings of the amplitude and time axes. This study is a starting point in employing the analysis of recurrences in investigations of a large class of real world signals.

Detecting errors in digital communications with CRC codes implemented with FPGA

Generally speaking, cyclic redundancy checks (CRCs) are used to detect errors from noise in digital data transmission. The technique is also sometimes applied to data storage devices, such as a disk drive. They also have been turned to verify the integrity of files in a system in order to prevent tampering and suggested as a possible algorithm for manipulation detection codes. It has been known that a CRC will not detect all errors but with random noise it is unlikely. In this paper, we present an efficient algorithm for parallel computation of the CRC in data transmission.

Fully integrated artificial intelligence solution for real time route tracking

In this paper the authors propose a solution in which an intelligent algorithm - genetic algorithm in our case - is used to generate commands for a robot in real time, so that the robot can determine the optimal moves considering several aspects: route tracking and low power consumption. Genetic algorithms are intelligent solutions for multi-criteria optimization and using them to find solutions to the optimization problems containing constrictions. However, they were designed as algorithms running on computer and therefore cannot ensure rapid generation of the solutions. On the other hand, the problem of determining the optimal response to command a robot requires real time response. The paper presents a method for hardware implementation and integration in a FPGA circuit of a genetic algorithm, in order to accelerate the convergence and to generate solutions in real time.

FPGA-based system for data acquisition and remote communication

In this paper we present a system that we made for acquisition and processing of data on multiple channels (8 in total) and transmitting the remote results through the Ethernet interface. The solution chosen by us involves the system implementing in FPGA where any processes running in parallel can be described - the acquisition processes from the AD sensors and possibly data processing processes. Our solution is a fully integrated system in an FPGA, which is composed of two main parts: the acquisition and processing of data part, made of modules (IP core) working in parallel and the communication part that consists of an architect of RISC processor, which takes the data processed by the parallel modules and can transmit them by running libraries for the TCPIP communication, just like a stand-alone processor. The solution was basically made on a Xilinx Spartan 6 FPGA circuit to which eight AD converters are connected, each taking data from one sensor. The model has an Ethernet interface which was used for remote data communication. On the processor implemented by us in FPGA we ran a TCP server, which through a client-server session allows the communication of the data.

Error detection and correction using LDPC in parallel Hopfield networks

Low Density Parity Check (LDPC) coding is a classical method for the detection and correction of errors. Its problem is the relatively complex operations which must be performed at encoding, and especially at decoding, to detect and correct errors caused by communication channels. We present a solution for the errors correction using regular LDPC and Hopfield network-based associative memories. Our solution solves this problem by using an associative memory based on the Hopfield network on the decoding stage, which stores the correct code words. This memory tends to transform the code words received with errors in errors free code words. We improved the ability of the associative memory by storing the code words in a specific format (pairs with their inverses, the constant number of “1” bits). These conditions are met by using LDPC coding. To be viable in communications, which require real-time correction, our proposed solution has used Hopfield networks operating in parallel, fully integrated into the Field Programmable Gates Array (FPGA) circuit.

Evaluation of microwave electromagnetic field absorbed by human thyroid gland

Radio communications is spreading more increasingly and people exposure to non-ionizing electromagnetic fields has increased due to emitting antennas. Human beings are complex electrochemical systems that communicate with the environment through electrical pluses. Electromagnetic field penetrates the human body and can alter functions of his cell, tissue and organs by modifying the biochemical processes. People have the ability to physiologically regulate their inner environment to ensure stability in response to fluctuations in the outside environment as a result of homeostasis process if the external sources are less than certain levels. The thyroid gland is one of the most exposed vital organs to electromagnetic radiation. Production of thyroid hormone is essential for metabolism, temperature regulation, and many other vital processes in the body. There are only few published papers reporting the effect of microwave radiation on thyroid. In this generic framework of the electromagnetic smog, that becomes less and less friendly, it is necessary to perform theoretical and experimental studies on the possible field levels that are currently applied to certain body parts. The aim of the present study was to assess the microwave field absorbed by human thyroid gland, and to predict local temperature increase for staff working in the vicinity of antennas transmitting high microwave power, for example on maritime ships. The temperature does not depend linear versus microwave power density. The finite element method was used to solve coupled equations governing the propagation microwave field in different environments with heat generation and propagation. In the paper are presented the necessary specific absorption rates, (SAR), as temperature increases up to 38–43 Celsius degrees for the frequency range between 1 and 4 GHz. Microwave signal power used by ships may lead to increase of 2–4 degrees Celsius.

VLSI implementation of an associative content addressable memory based on Hopfield network model

The content addressable memory (CAM) is allowed to search a data word without knowing where its address is. In addition, it is permissible to associate the content of the location or neighboring locations where the data word was identified. This paper presents our own approach for VLSI hardware implementation of the CAM memory. The proposed solution uses a Hopfield neural network model and is characterized by simplicity and the possibility of using the same hardware structures for saving may data patterns. Will be presented design methods and implementation to VLSI circuit structures, the performance of our solution and experimental results.

Optimization of sigma-delta modulator based on artificial immune algorithms

The artificial immune algorithms are implemented by the human immune system. They inspired a technique that has the ability to find optimal solutions in a nonlinear search space. This kind of technique increases the computing speed and converges faster than the genetic algorithms. In this paper is presented an application of artificial immune algorithms in order to optimize the transfer function of a sigma delta modulator. It is also presented the base method which can perform the adjustment coefficients, in decimal, of the sigma delta modulator. The results are compared with those obtained by the method of optimization based on both mathematical and genetic algorithms.

Real time system for instrumental sound extraction and recognition

This article presents a “smart sound sensor” solution. More specifically it is a system on a chip capable of extracting sound patterns from an input signal captured by a microphone, patterns that correspond to certain musical instruments. The aim is to identify sounds like: voices, ambient sounds or musical sounds. Such applications generally involve the presence of a computer that analyzes offline the audio sequence. There are also smartphone security applications that rely on voice recognition templates that were previously stored and which work by processing the input sequence offline, after recording. Unlike these methods, our solution enables online real time identification of a group of musical instruments and allows extracting of certain sound sequences that belong to a particular instrument so that it can be isolated for sole playing. The solution was implemented in an integrated system “sound sensor” that analyzes the input signal in real time and can give output data without additional elements.