Adaptive Blind Signal and Image Processing

In this paper we focus on Bayesian blind and semi-blind adaptive signal processing based on a broadband MIMO FIR model (e.g., for blind source separation (BSS) and blind system identification (BSI)). Specifically, we study in this paper a framework allowing us to systematically incorporate various types of prior knowledge: (1) source signal statistics, (2) deterministic knowledge on the mixing system, and (3) stochastic knowledge on the mixing system. In order to exploit all possible types of source signal statistics (1), our considerations are based on TRINICON, a previously introduced generic framework for broadband blind (and semi-blind) adaptive MIMO signal processing. The motivation for this paper is threefold: (a) the extension of TRINICON to Bayesian point estimation to address (3) in addition to (1), and (b) more specifically to unify system-based blind adaptive MIMO signal processing with the tracking of time-varying scenarios, and finally (c) to show how the Bayesian TRINICON-based tracking can be formulated as a sequence estimation approach on arbitrary partly smooth manifolds. As we will see in this paper, the Bayesian approach to incorporate stochastic priors and the manifold learning approach to exploit deterministic system knowledge (2) complement one another very efficiently in the context of TRINICON.

This chapter is focused on spatially adaptive image processing for color images in the context of the General Adaptive Neighborhood Image Processing (GANIP) approach. The GANIP was first defined for gray-tone images and is here extended to color images. A set of local adaptive neighborhoods is defined for each image point, depending on the color intensity function of the image. These adaptive neighborhoods are then used as spatially adaptive operational windows for defining adaptive Choquet filters and adaptive morphological filters. The resulting adaptive operators are successfully applied and compared with the classical operators for image restoration, enhancement and segmentation of color images.

Over the last 30 years, adaptive digital signal processing has progressed from being a topic for a strictly graduate-level advanced class in signal processing theory to one that is part of the core curriculum for many undergraduate classes in signal processing. The key reason is the continued advance of communications technology, with its need for echo control and equalisation, and the widespread use of adaptive filters in audio, biomedical, and control applications. This paper reviews the basic theory and applications of adaptive signal processing and also presents information on multimedia teaching methods that are being used on a 24 hour course module on adaptive signal processing. In particular it presents some recent World Wide Web Java programs that were developed for teaching and self-study. Information is provided on how to download and run the Java applet.

The possibilities of improving electromagnetic compatibility in mobile communication systems of next generations are analysed based on application of adaptive space-time signal processing in antennas of base stations. Features of solving the problem of interference suppression in adaptive antenna arrays are considered using recurrent adaptation algorithms realized on the basis of the minimum mean square error criterion. The results of modelling adaptation of signal processing and its analysis are presented. A promising method of interference suppression using polarisation-time processing of signals using optimal stochastic control is considered.

Potential areas of development in optically based adaptive signal processing are considered together with an assessment of current technological directions. The rapid adoption of superhigh‐throughput lightwave communications media and systems suggests that fully optical signal processing would be very desirable to take advantage of the speed. The broad conclusion of this analysis is that at present it is not feasible to implement fully optical adaptive signal processors but that in time it might be so. ‘Not yet’ is our surmise. A brief literature survey and description of existing prototype optoelectronic adaptive systems is also provided.

Two new methods developed to improve the classical preset time count rate meters by using the adaptive signal processing tools, are presented. An optimized detection algorithm that senses the change of the mean count rate was implemented in both methods. Three low-pass filters of various structures with adaptive parameters to implement the control of the mean count rate error were considered. An adaptation algorithm for preset time interval calculations was devised and implemented in both methods differing only in the place of its execution (before or after the low-pass filters). The simulated and realized methods, using the developed algorithms, guarantee: the response time not in excess of 2s for the mean count rate higher than 2c/s and the controllable mean count rate error in the range of ±4% to ±10%.

We considered the issues of synthesis of the algorithms for adaptive nonlinear signal processing using feed-forward blocks of nonlinear transformation under the influence of non-Gaussian noise with unknown density of distribution of instantaneous values or its envelope. It is shown that to plot the adaptive feed-forward blocks of nonlinear transformation, the algorithms for estimating the parameters of linear model of probability density function of noise can be used. This model is presented in the form of a generalized polynomial of decomposition in a series of linearly independent functions, and, also, in the form of nonlinear models, such as generalized Gaussian distribution and abnormally cluttered distribution.

Two optimized adaptive digital signal processors were implemented to the preset time count rate meters. Three mean count rate meters, based on the developed adaptive digital signal processors, were realized and were used for experimental validation of proposed adaptive digital signal processors. The experimental results, conducted without and with radioisotope for the specified errors of 10% and 5%, showed to agree well with theoretical predictions.

As an attempt to introduce interactive, content‐dependent adaptive (ICDA) image processing, a simple but powerful active image sensing and two image enhancement methods are introduced via adaptive CNN‐UM sensor‐computers. Thus, the method ICDA can be used for adaptive control of image sensing and for subsequent on‐line or off‐line image enhancement as well. The algorithms use both intensity and contrast content. The image sensing technology can be realized with the current CNN‐UM chip. Our first image enhancement method is also executable on this chip, but it is more suitable for the adaptive cellular neural network universal machine (ACNN‐UM) architecture. Some results of simulator and chip experiments and an adaptive extended cell are presented. Our second, dynamical image enhancement method is planned to be executable on a multi‐layer, complex cell CNN architecture. In (Proceedings of the 6th IEEE International Workshop on Cellular Neural Networks and their Applications(CNNA‐2000) Catania, 2000; 213–217) 3‐layer architecture is described which is capable of realizing the main part of the second enhancement method.The main issues of our paper are as follows: the novel outlook of the ICDA framework, three new methods for two key application areas of CNN‐UM, the notion of ‘regional’ adaptive computing, the novelty of application of equilibrium‐computing in the third method. However, the key novelty of our work is not just a new method and a new realization: by combining sensing and computing, dynamically and pixelwise, a new quality becomes practical. Copyright © 2002 John Wiley & Sons, Ltd.

The increasing demand for intelligent surveillance systems has led to the integration of Internet of Things (IoT) technology with advanced image processing techniques to enhance security applications. This paper presents a novel framework that combines IoT-enabled devices with adaptive image processing, leveraging the power of edge computing and machine learning to optimize real-time analysis of visual data. Traditional surveillance systems often face limitations such as high latency, bandwidth constraints, and inadequate adaptability to dynamic environments. Our proposed solution addresses these challenges by processing data closer to the source, reducing latency and bandwidth usage while maintaining high processing accuracy. The framework integrates adaptive algorithms for dynamic resolution adjustment, context-aware object detection, and real-time anomaly detection, allowing the system to intelligently adjust based on scene complexity, environmental conditions, and available computational resources. Additionally, the use of machine learning techniques enhances the system’s ability to learn from new patterns and adapt to evolving security threats. Experimental evaluations demonstrate the proposed system’s superior performance in terms of object detection accuracy, processing speed, and energy efficiency compared to traditional centralized surveillance systems. This paper also discusses the implementation of secure communication protocols and privacy-preserving techniques to ensure data security in IoT environments. The proposed framework offers a scalable and flexible solution, making it suitable for a wide range of security applications, from public spaces to private facilities. Future work will focus on improving the system's robustness and extending its capabilities for large-scale deployments.

A new adaptive algorithm for system identification and adaptive signal processing has been proposed in this paper. The algorithm can be viewed as a combination of the output-error type and the equation-error type of identifier (that is, a combination of the IIR adaptive filter and the FIR adaptive filter in signal-processing applications). It has been shown both mathematically and by computer simulation that the proposed algorithm can guarantee the globally asymptotical stability. The transient behaviour of the algorithm is better than the conventional output-error type of identifier, especially in the case where the system (signal model) has its eigenvalues on the unit circle. Under certain conditions on the system model (signal model), no moving average (MA) smoothing of the generalized error is needed for the strictly positive realness (SPR) conditions.

Impact induced damage assessment by means of Lamb wave image processing

Faced with complex electromagnetic environment and varieties of adaptive radar waveforms, radar signal analysis and identification becomes more and more complex. Considering two important physical quantities - time and frequency in modern signal processing methods, this paper proposes that the joint time-frequency analysis (JTFA) method based on fractional Fourier transform (FrFT) and short-time Fourier transform (STFT) is applied to adaptive radar signal processing. The simulation results show that the joint time-frequency analysis method is superior to single short-time Fourier transform, getting a better analysis of results. The joint time-frequency analysis method provides the joint distribution of the time domain and frequency domain for adaptive radar signal analysis and describes the relationship between signal frequency and time.

The FDSP (Fuzzy Digital Signal Processor) is a new device that combines fuzzy processor and DSP in a single core well adapted to be hosted by low cost microcontrollers. The FDSP offers economic solution to implement adaptive fuzzy digital signal processing. The fuzzy processor acts as a controller or an adaptive element to a DSP system to implement adaptive filters, noise cancellers, estimators, channel equalizers, etc. We demonstrate the use of the FDSP to implement hard disk drive controller that combines quasi-optimum seek-track control, adaptive noise filters, and velocity estimator. Other applications of the FDSP include non-linear imaging filters and motion estimators and other systems for multimedia processing.

Includes original methods developed for solution of the problems analysed in previous chapters. Firstly, the main approaches are proposed oriented on development of effective algorithms of adaptive spatial signal processing with not exactly known parameters. The probabilistic approach for synthesis of robust algorithms of ASSP is analysed. Next, the deterministic approach for synthesis of robust algorithms of ASSP is represented as a solution of some modified optimization problems. Development of methods based on linear and correlation restrictions. The third part is devoted to development of robust algorithms of ASSP. The last part proposes a projection approach for development of effective algorithms of adaptive spatial signal processing with not exactly known parameters.KeywordsAntenna ArrayDirectivity CharacteristicDeterministic ApproachAntenna ElementOptimization TaskThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.