Dynamic Infrastructural Distributed Denial of Service (I-DDoS) attacks constantly change attack vectors to congest core backhaul links and disrupt critical network availability while evading end-system defenses. To effectively counter these highly dynamic attacks, defense mechanisms need to exhibit adaptive decision strategies for real-time mitigation. This paper presents a novel Autonomous DDoS Defense framework that employs model-based reinforcement agents. The framework continuously learns attack strategies, predicts attack actions, and dynamically determines the optimal composition of defense tactics such as filtering, limiting, and rerouting for flow diversion. Our contributions include extending the underlying formulation of the Markov Decision Process (MDP) to address simultaneous DDoS attack and defense behavior, and accounting for environmental uncertainties. We also propose a fine-grained action mitigation approach robust to classification inaccuracies in Intrusion Detection Systems (IDS). Additionally, our reinforcement learning model demonstrates resilience against evasion and deceptive attacks. Evaluation experiments using real-world and simulated DDoS traces demonstrate that our autonomous defense framework ensures the delivery of approximately 96 – 98% of benign traffic despite the diverse range of attack strategies.

This paper proposed a method of online non-parameter identification of nonlinear ship motion systems. Firstly, we use Mariner to generate a certain amount of ship motion data to train the LWPR model. Then the ship travels along a set track. During this process, the sensors continuously obtain the distance, radial velocity and azimuth of the ship relative to the ship, and then completes the construction of simulation data. Next, the performance of the algorithm is verified which uses the Kalman filtering framework. Finally, the estimated value is further used for updating the LWPR model to achieve the purpose of online learning, and the updated model will be used for the next prediction. The experimental results show that the online modeling and tracking method proposed in this paper has higher tracking accuracy than the parameter estimation techniques.

Auditory, which served as one of the five sense systems, play a vital role in human beings’ daily life. Among the many auditory detection techniques, Auditory Brainstem Response (ABR) is widely chosen and studied for its convenience and objectivity. The averaging (Ave) technique is the currently applied method to extract ABR from the EEG signals and is regarded as the gold standard in the clinic. However, the Ave technique is not suitable for noisy condition, like active behavioral condition, which requires the subjects to keep stay during the whole ABR test and is therefore not suitable for newborn. To extract ABR signals from the real active behavioral condition, an adaptive kalman filter (AKF) technique was proposed and systematically investigated from the morphology aspect in two conditions, namely rest and active behavioral (chewing) conditions. The results showed that in rest condition, the ABR signal obtained by the AKF method was highly similar to that of the gold standard method, and the latencies and amplitudes of characteristics waves were also alike. Moreover, we analyzed the latencies and amplitudes of the characteristics waves and CC between the standard ABR and the different method-based ABR. The analyses showed that the AKF had the potential on the extraction of ABR in active behavioral condition. The AKF method provides a new way to robust denoise, and opens a window for ABR acquisition in active behavioral condition, making ABR acquisition in daily life more possible.

Entering the critical year of the 14th Five Year Plan, China s information security industry has entered a new stage of development. With the increasing importance of information security, its industrial development has been paid attention to, but the data fragmentation of China s information security industry is serious, and there are few corresponding summaries and predictions. To achieve the development prediction of the industry, this article studies the intelligent prediction of information security industry data based on machine learning and new adaptive weighted fusion, and deduces the system based on the research results to promote industry development. Firstly, collect, filter, integrate, and preprocess industry data. Based on the characteristics of the data, machine learning algorithms such as linear regression, ridge regression, logical regression, polynomial regression and random forest are selected to predict the data, and the corresponding optimal parameters are found and set in the model creation. And an improved adaptive weighted fusion model based on model prediction performance was proposed. Its principle is to adaptively select the model with the lowest mean square error (MSE) value for fusion based on the real-time prediction performance of multiple machine learning models, and its weight is also calculated adaptively to improve prediction accuracy. Secondly, using technologies such as Matplotlib and Pyecharts to visualize the data and predicted results, it was found that the development trend of the information security industry is closely related to factors such as national information security laws and regulations, the situation between countries, and social emergencies. According to the predicted results of the data, it is observed that both industry input and output have shown an upward trend in recent years. In the future, China s information security industry is expected to maintain stable and rapid growth driven by the domestic market.

Spatial field digital modulation (SFDM) communication system is a special index modulation (IM) technique with low hardware complexity and physical layer security potential. However, the deployment of the SFDM system is always complicated and time-consuming. To solve the problems, an adaptive SFDM system without phase measurement is proposed and implemented in this paper. We design a system architecture fit for self-adjustment and propose the corresponding adaptive algorithm. The state isolation and the BER performance are measured under an indoor channel, which verifies its validity.

This paper investigates the output feedback security control problem of switched nonlinear systems (SNSs) against denial-of-service (DoS) attacks. A novel switched observer-based neural network (NN) adaptive control algorithm is established, which guarantees that all the signals in the closed-loop system remain bounded. Note that when a DoS attacker is active in the Sensor-Controller channel, the controller cannot acquire accurate information, which leads to the standard backstepping technique not being workable. A set of NN adaptive switching-like observers is designed to tackle the obstacle for each subsystem. Further, by combining the proposed observer with the backstepping technique, an NN adaptive controller is constructed and the dynamic surface control method is borrowed to surmount the complexity explosion phenomenon. Finally, an illustrative example is provided to demonstrate the effectiveness of the proposed control algorithm.

Adaptive security is considered as an approach in cybersecurity that analyzes events and against events and behaviors to protect a network. This study will provide details about the different algorithms being used to secure networks. These approaches are driven by a small quantity of labeled data and a massive amount of unlabeled data. In this context, contemporary semi-supervised learning strategies base their operations on the assumption that the distributions of labeled and unlabeled data are comparable. This assumption has a substantial influence on how well these strategies perform overall. If unlabeled data contain information that does not belong to a particular category, the efficiency of the system will deteriorate.

Forecasting technology plays an important role in the construction of systems for detecting anomalies in dynamic data flows of automated process control systems (APCS) resulting from the impact of cyberattacks. To form a forecast of the studied signals, methods for forming a single-component forecast and a multi-component forecast of an information signal using a linear prediction digital filter are considered. It is shown that for the detection of anomalies in the observed signals of APCS, the predictor prediction error signal implemented using a linear prediction filter is informative. The high information content of the use of spectral analysis of the prediction error signal in detecting anomalies in the observed signals of automated process control systems is shown.

Despite various distributed denial-of-service (DDoS) filtering solutions proposed and deployed throughout the Internet, DDoS attacks continue to evolve and successfully overwhelm the victims with DDoS traffic. While current DDoS solutions in general employ a fixed filtering granularity (e.g., IP address, 4-tuple flow, or service requests) with a specific goal (e.g., maximum coverage of DDoS traffic), in this paper we investigate adaptive DDoS filtering. We design and experiment algorithms that can generate and deploy DDoS-filtering rules that not only adapt to the most suitable and effective filtering granularity (e.g., IP source address and a port number vs. an individual IP address vs. IP prefixes at different lengths), but also adapt to the first priorities of victims (e.g., maximum coverage of DDoS traffic vs. minimum collateral damage from dropping legitimate traffic vs. minimum number of rules). We evaluated our approach through both large-scale simulations based on real-world DDoS attack traces and pilot studies. Our evaluations confirm that our algorithms can generate rules that adapt to every distinct filtering objective and achieve optimal results.

The low-frequency radiated sound field can be effectively controlled through the adaptive active control method in theory. However, its application in underwater radiated noise control is not wide. In the active control system, especially the multi-channel feedback system, the step size has a very tremendous influence on the performance of the adaptive filter. If the step size is set unreasonably, the calculation results will not converge. The appropriate step size varies from case to case. For simple cases, the empirical value can be adopted to set the step size. When the numerical difference between channels is large, and when the control physical quantity such as sound pressure changes greatly with time, an determined step length can t meet the control requirements. In particular, it is difficult to choose the step size when the accurate reference signal cannot be obtained. The application of adaptive active methods in underwater noise control is limited to some extent by this problem. To solve this problem, this essay carried out the research of Filtered-X Least Mean Squares (FxLMS) algorithm based on variable step size, and carried out the corresponding numerical analysis and pool experiment to verify the feasibility of applying to underwater noise control.

Wave filtering is one of the mandatory features of the state estimators in a dynamic position system. The optimization of statistical parameters of these state estimators can be done by covariance matching algorithms and appropriate objective (cost) functions. The proposed cost function has predictive behavior, based on some tuning parameters, which control the quality of wave filtering. These parameters assure convergence of the solution and consistent results in different adaptive algorithms based on the Kalman filter framework as AKF, AEKF, and AUKF.

An adaptive command filter control of nonlinear system with friction input is formulated in this paper. First, based on the obtained state space model, a command filter control method is proposed, which can address the “explosion of complexity” problem existed in traditional backstepping design and ensure the asymptotic convergence of the tracking errors. Moreover, to cope with the problem of filter error between filter output and virtual control signal, dynamic error compensation system is designed. Next, a HONN system is employed to simplify the calculation and approximate the uncertainties in the system. At last, in order to clarify the effectiveness of the above theory, simulation results are given.

To solve the problem that the filtering accuracy of the online calibration decreases or even diverges due to timevarying noise and outlier value interference, a SINS/CNS/GPS high-precision integrated navigation online calibration method based on the improved Sage-Husa adaptive filtering algorithm is designed. In the proposed method, a 21-dimensional state space model and 9-dimensional measurement model are established. Furthermore, on the basis of the simplified Sage-Husa adaptive filtering algorithm, a smoothing estimator and an adaptive robust factor are introduced to suppress the influence on the filtering accuracy due to the abnormal disturbances in the measurement information, which improving the online calibration accuracy of integrated navigation. The simulation results show that the online calibration method based on the improved Sage-Husa adaptive filtering algorithm can better calibrate the error parameters, especially the calibration of the lever arm error for the east and up directions.

This article presents a new concept of fully analogue adaptive filters. The adaptation is based on fully analogue neural networks. With the use of a filter bank, it can be used for high frequency and real-time adaptation. The properties of this concept are verified using electronic circuit simulations.

At the heart of most adaptive filtering techniques lies an iterative statistical optimisation process. These techniques typically depend on adaptation gains, which are scalar parameters that must reside within a region determined by the input signal statistics to achieve convergence. This manuscript revisits the paradigm of determining near-optimal adaptation gains in adaptive learning and filtering techniques. The adaptation gain is considered as a matrix that is learned from the relation between input signal and filtering error. The matrix formulation allows adequate degrees of freedom for near-optimal adaptation, while the learning procedure allows the adaption gain to be formulated even in cases where the statistics of the input signal are not precisely known.

This paper considers adaptive signal equalization in a channel with inter-symbol interference (ISI) distortion. In this process two adaptive FIR filters with different "forgetting factors" are used to update their filter coefficients. RLS algorithm is applied to optimize the filter coefficients. A comparison to a Genetic algorithm was done. The difference between the estimated mean square errors of both filters provides indication on how to change the "forgetting factors" to get closer to their optimal value. Finally, the computational efforts for filters with five different filter orders are compared. The obtained results prove the applicability of the presented approach.

Electrocardiography (ECG) is the most popular non-invasive method for generating an Electrocardiogram which contains some very interesting information about the electrical and myographic activities of the heart. It is a graph of voltage vs. time of the electrical activity of the heart using electrodes connected on the skin in various configurations. Due to the noninvasive nature of ECG and also due to capacitive or inductive coupling in this electrical circuit for ECG acquisition or electromyographic noises due to muscles adjacent to heart there is usually significant noises present in a typical ECG which makes it harder to analyze. There are many methods for denoising ECGs. In this paper an adaptive unscented Kalman filter, where the measurement noise covariance matrix is varied adaptively, is used for denoising acquired discrete ECG signals. The filtered output as well as the improvement of SNR is compared with other existing denoising frameworks like discrete wavelet transform and digital filters and extended Kalman filter, and unscented Kalman filter. The Adaptive Unscented Kalman Filter performed better than the aforementioned existing filtering algorithms in terms of maximum output SNR and MSE computed using Monte Carlo simulation.

This paper proposes improved combined step-size sign subband adaptive filter (ICSS-SSAF) algorithms with variable mixing factors robust to non-Gaussian noises such as impulsive noise. The CSS scheme is adopted to resolve a trade-off problem of step size in the SSAF, combining two adaptive filters with a large step size for a fast convergence rate and a small step size for low steady-state misalignment. Variable mixing factors (VMFs), whose values are changed at every iteration, are introduced to combine the two adaptive filters. To design the VMFs, a modified sigmoidal or arctangent function is employed. They are updated indirectly to minimize the power of approximated system output error, unlike the conventional algorithm using the 1 norm of the error vector composed of error signals divided by subbands. The recursive forms of VMFs are acquired by adopting the gradient method. The simulation results show that the proposed algorithms perform better than conventional algorithms in system identification scenarios.

Noise has become a significant concern in every domain. For instance, in image processing, we can see background noise when we take a snap, also in the field of communication, the information is corrupted by the noises present in the environment, and at the time of decryption, it is becoming challenging. Back then, in earlier days, discrete filters that had fixed frequency response were used to minimize the level of Noise in the information signals. But these filters were not effective as most noise sources have a flat wideband spectrum. After the availability of digital signal processors, to obliterate the wideband Noise, adaptive filters are frequently used in communication systems and digital signal processing systems to filter noisy signals. The Adaptive Noise Cancellation (ANC) approach helps to eliminate the Noise by altering its transient parameters dependent on the incoming signal. In this article, the performance of LMS, NLMS and RLS algorithms is studied for various types of ambient noises. A speech signal that is corrupted by engine noise, waterfall noise, and audio noise and with echo are applied to an ANC filter and the improvement in signal to noise ratio is evaluated with different adaptive filter algorithms.

Through the thorough exploration of the adaptive filter structure and the LMS adaptive filter algorithm, the filter performance of the adaptive filter algorithm can be clearly mastered.The solution formula of LMS algorithm is based on it, and DSP software programming and Matlab simulation programming methods are used to lay the foundation for the effective implementation of LMS algorithm.Therefore, based on the adaptive filtering algorithm, the embedded software simulation development system is analyzed to help the application of adaptive filtering theory.

Existing defense strategies against adversarial attacks (AAs) on AI/ML are primarily focused on examining the input data streams using a wide variety of filtering techniques. For instance, input filters are used to remove noisy, misleading, and out-of-class inputs along with a variety of attacks on learning systems. However, a single filter may not be able to detect all types of AAs. To address this issue, in the current work, we propose a robust, transferable, distribution-independent, and cross-domain supported framework for selecting Adaptive Filter Ensembles (AFEs) to minimize the impact of data poisoning on learning systems. The optimal filter ensembles are determined through a Multi-Objective Bi-Level Programming Problem (MOBLPP) that provides a subset of diverse filter sequences, each exhibiting fair detection accuracy. The proposed framework of AFE is trained to model the pristine data distribution to identify the corrupted inputs and converges to the optimal AFE without vanishing gradients and mode collapses irrespective of input data distributions. We presented preliminary experiments to show the proposed defense outperforms the existing defenses in terms of robustness and accuracy.