State-space Model Research Articles

Spatial-spectral attention state-space fusion network for tree carbon estimation using multi-source data

Abstract Amid rapid industrialization and urbanization, global warming and the urban heat island effect are intensifying. There is an urgent need for rapid and accurate monitoring of carbon storage in urban vegetation to assess and mitigate these environmental issues. While methods exist for estimating above-ground carbon (AGC) stocks in individual trees using deep learning, fewer studies integrate multi-source remote sensing data with these techniques. This study introduces a network that leverages spatial-spectral attention and state-space models, called the spatial-spectral attention state-space fusion network (S4FNet). S4FNet utilizes UAV hyperspectral imaging (HSI) data, light detection and ranging (LiDAR) data, and RGB data to estimate AGC storage in individual trees. This study conducted experiments at two sites, the South and North areas of Shenzhen University’s Yuehai Campus, Shenzhen City, Guangdong Province, China, and compared S4FNet with eight traditional machine learning methods and a deep learning method. The experimental results demonstrate that S4FNet achieves the best performance across all metrics, confirming its excellent capability in predicting AGC stocks using multi-source remote sensing data.

Heterogeneously Networked Evolutionary Games With Intergroup Conflicts.

Network games primarily explore the intricacies of individual interactions and adaptive strategies within a network. Building upon this framework, the present study delves into the modeling, analysis, and control of heterogeneously networked evolutionary games with intergroup conflict (HNEG-IC), where attacking players possess area-monitoring capabilities with limited attacking power. To begin with, a mathematical model is introduced to capture intragroup strategy dynamics and intergroup conflicts of HNEGs-IC via the algebraic state space representation (ASSR). A necessary and sufficient condition for achieving global cooperation of HNEGs-IC is established. Then, a criterion for verifying the κ -cooperation below a certain mortality is presented. Considering the HNEGs-IC with strategy feedback control, it is proven that the feedback control, subject to global cooperation, is robust to conflicts when the intersection of the strategy threshold set and the reachable set of the preset initial strategy profiles is empty. Finally, for verification and demonstration, the obtained results are applied to a simplified virtual game model of the NATO and the Warsaw Pact.

Data-Driven Modeling of DC–DC Power Converters

This article presents a data-driven methodology for modeling DC–DC power electronic converters. Using the proposed methodology, the dynamics of a converter can be captured, thereby eliminating the need for explicit theoretical modeling methods. This approach only requires the acquisition of fundamental measurements: currents through inductors and voltages across capacitors. The acquired data are used to construct a linear difference system, which is algebraically manipulated to form a state–space representation of the converter under analysis. Three DC–DC converter topologies were analyzed, and their resulting models were tested and compared with simulation data, yielding an average error deviation of approximately 2% for current signals and 4% for voltage signals, demonstrating precise tracking of the actual dynamics. The proposed data-driven methodology could simplify the implementation of adaptive control strategies in larger-scale solutions or in the interconnection of multiple converters.

IFNet: Data-driven multisensor estimate fusion with unknown correlation in sensor measurement noises

In recent years, multisensor fusion for state estimation has gained considerable attention. The effectiveness of the optimal fusion estimation method heavily relies on the correlation among sensor measurement noises. To enhance estimate fusion performance by mining unknown correlation in the data, this paper introduces a novel multisensor fusion approach using an information filtering neural network (IFNet) for discrete-time nonlinear state space models with cross-correlated measurement noises. The method presents three notable advantages: First, it offers a data-driven perspective to tackle uncertain correlation in multisensor estimate fusion while preserving the interpretability of the information filtering. Second, by harnessing the RNN’s capability to manage data streams, it can dynamically update the fusion weights between sensors to improve fusion accuracy. Third, this method has a lower complexity than the state-of-the-art KalmanNet measurement fusion method when dealing with the fusion problem involving a large number of sensors. Numerical simulations demonstrate that IFNet exhibits better fusion accuracy than traditional filtering methods and KalmanNet fusion filtering when correlation among measurement noises is unknown.

On the axiomatization of an unawareness structure from knowing-whether operators

This paper shows that, on a generalized state space model of unawareness, an agent’s underlying knowledge is axiomatized from her knowing-whether operator if and only if her knowledge satisfies the Truth Axiom: whenever the agent knows an event, the event holds. The agent knows whether an event obtains if she knows it or knows its negation. Different knowledge operators lead to different knowing-whether operators if knowledge is truthful. Conversely, for any knowing-whether operator, there is a unique truthful knowledge operator that induces the given knowing-whether operator: the agent knows an event if and only if she knows whether the event holds and the event indeed holds. Qualitative or probabilistic beliefs may not be recovered from believing-whether. This paper then axiomatizes properties of knowledge and common knowledge, in terms of knowing-whether. The main contributions of the paper are as follows. First, conceptually, this paper provides a generalized-state-space model of knowledge and unawareness in which the only assumption on knowledge is the Truth Axiom. Second, practically, this paper may provide a simple way to construct a generalized-state-space model.

Proportional Integral Type Sliding Function For DC-DC Boost Converter In Solar Power System

Background: DC-DC Boost Converters are crucial components in solar power systems, requiring effective control for optimal performance. This paper introduces an adaptive sliding mode controller with a modified sliding function to overcome the limitations of classic sliding mode control in DC-DC boost converters. Materials and Methods: The study presents a detailed mathematical model of the DC-DC boost converter and derives a state-space representation. A Proportional-Integral (PI) type sliding function is proposed, introducing an additional tuning parameter γ that can be adaptively adjusted based on load variations. The control law and existence conditions for sliding modes are established, demonstrating the robustness of the proposed method to load changes. Results: Simulation results compare the performance of the proposed PI-type sliding mode control (SMC) with classic SMC for a DC-DC boost converter under varying load conditions. The PI-type SMC demonstrates superior dynamic performance, recovering from a sudden load change in 10 ms compared to 30 ms for classic SMC. Conclusion: The proposed PI-Type Sliding Function for DC-DC Boost Converters in solar power systems offers improved adaptability to load variations and faster dynamic response, potentially enhancing the efficiency and reliability of solar power systems.

Bayesian-Learning-Based Diffusion Least Mean Square Algorithms Over Networks.

To improve the learning performance of the conventional diffusion least mean square (DLMS) algorithms, this article proposes Bayesian-learning-based DLMS (BL-DLMS) algorithms. First, the proposed BL-DLMS algorithms are inferred from a Gaussian state-space model-based Bayesian learning perspective. By performing Bayesian inference in the given Gaussian state-space model, a variable step-size and an estimation of the uncertainty of information of interest at each node are obtained for the proposed BL-DLMS algorithms. Next, a control method at each node is designed to improve the tracking performance of the proposed BL-DLMS algorithms in the sudden change scenario. Then, a lower bound on the variable step-size of each node of the proposed BL-DLMS algorithms is derived to maintain the optimal steady-state performance in the nonstationary scenario (unknown parameter vector of interest is time-varying). Afterward, the mean stability and the transient and steady-state mean square performance of the proposed BL-DLMS algorithms are analyzed in the nonstationary scenario. In addition, two Bayesian-learning-based diffusion bias-compensated LMS algorithms are proposed to handle the noisy inputs. Finally, the superior learning performance of the proposed learning algorithms is verified by numerical simulations, and the simulated results are in good agreement with the theoretical results.

Rao–Blackwellized Particle Filter Using Noise Adaptive Kalman Filter for Fully Mixing State-Space Models

Rao–Blackwellized Particle Filter Using Noise Adaptive Kalman Filter for Fully Mixing State-Space Models

Explicit and implicit locomotor learning in individuals with chronic hemiparetic stroke.

Motor learning involves both explicit and implicit processes that are fundamental for acquiring and adapting complex motor skills. However, stroke may damage the neural substrates underlying explicit and/or implicit learning, leading to deficits in overall motor performance. Although both learning processes are typically used in concert in daily life and rehabilitation, no gait studies have determined how these processes function together after stroke when tested during a task that elicits dissociable contributions from both. Here, we compared explicit and implicit locomotor learning in individuals with chronic stroke to age- and sex-matched neurologically intact controls. We assessed implicit learning using split-belt adaptation (where two treadmill belts move at different speeds). We assessed explicit learning (i.e., strategy-use) using visual feedback during split-belt walking to help individuals explicitly correct for step length errors created by the split-belts. After the first 40 strides of split-belt walking, we removed the visual feedback and instructed individuals to walk comfortably, a manipulation intended to minimize contributions from explicit learning. We used a multirate state-space model to characterize individual explicit and implicit process contributions to overall behavioral change. The computational and behavioral analyses revealed that, compared with controls, individuals with chronic stroke demonstrated deficits in both explicit and implicit contributions to locomotor learning, a result that runs counter to prior work testing each process individually during gait. Since poststroke locomotor rehabilitation involves interventions that rely on both explicit and implicit motor learning, future work should determine how locomotor rehabilitation interventions can be structured to optimize overall motor learning. NEW & NOTEWORTHY Motor learning involves both implicit and explicit processes, the underlying neural substrates of which could be damaged after stroke. Although both learning processes are typically used in concert in daily life and rehabilitation, no gait studies have determined how these processes function together after stroke. Using a locomotor task that elicits dissociable contributions from both processes and computational modeling, we found evidence that chronic stroke causes deficits in both explicit and implicit locomotor learning.

Interaction between three key species in the sea ice-reduced Arctic Barents Sea system.

Population dynamics depend on trophic interactions that are affected by climate change. The rise in sea temperature is associated with the disappearance of sea ice in the Arctic. In the Arctic part of the Barents Sea, Atlantic cod, capelin and polar cod are three fish populations that interact and are confronted with climate-induced sea ice reductions. The first is a major predator in the system, while the last two are key species in Arctic and sub-Arctic ecosystems, respectively. There are still many unknowns regarding how predicted environmental change may influence the joint dynamics of these populations. Using time series from a 32 year long survey, we developed a state-space model that jointly modelled the dynamics of cod, capelin and polar cod. Using a hindcast scenario approach, we projected the effect of reduced sea ice on these populations. We show that the impact of sea ice reduction and concomitant sea temperature increase may lead to a decrease of polar cod abundance at the benefit of capelin but not of cod which may decrease, resulting in strong changes in the food web. Our analyses show that climate change in the Arcto-boreal system can generate different species assemblages and new trophic interactions, which is the knowledge needed for effective management measures.

Estimating the impact of sarcoptic mange epidemic on the population size of wild raccoon dogs (Nyctereutes procyonoides) from wildlife rescue data

The impact of infectious diseases on host populations is often not quantified because it is difficult to observe the host population and infectious disease dynamics. To address this problem, we developed a state-space model to simultaneously estimate host population and disease dynamics using wildlife rescue data. Using this model, we aimed to quantify the impact of sarcoptic mange on a Japanese raccoon dog population by estimating the change in their relative population size. We classified the status of rescued raccoon dogs into four categories: i) rescued due to infection with mange, ii) rescued due to traffic accidents without mange, iii) rescued due to traffic accidents with mange, and iv) rescued due to causes other than traffic accidents or mange. We modelled the observation process for each category and fitted the model to the reported number of raccoon dogs rescued between 1990 and 2010 at three wildlife rescue facilities in Kanagawa Prefecture, Japan. The mortality rate induced by mange was estimated to be 1.09 (95% credible interval (CI): 0.47–1.72) per year. The estimated prevalence of sarcoptic mange ranged between 4–80% in the study period. When a substantial prevalence of mange was observed (1995 to 2002), the host population size decreased by 91.2% (95% credible intervals: 86.3–94.7). We show that the impact of infectious disease outbreak on the wildlife population can be estimated from the time-series data of wildlife rescue events due to multiple causes. Our estimates suggest that sarcoptic mange triggered a substantial decrease in the Japanese wild raccoon dog populations.

Bayesian State-space modelling of stock Markets in G7 Countries

This work examines the impact of Coronavirus disease (COVID-19) on the stock market of Group of Seven (G7) countries during the first wave of COVID-19 using daily data from March, 1st of 2020 to December, 31st of 2020. Focussing on such period, a Bayesian Structural Time Series Model (BSTSM) was used to capture the effects of first wave of COVID-19 on the stock market performance of these G7 countries by employing a Markov Chain Monte Carlo (MCMC) method. We considered an Autoregressive (AR) model with time-varying parameters and a local linear trend model to know if the stock price of these countries during the period of the first wave of COVID-19 is changing over time. There was a stochastic trend in stock prices of G7 countries during the period of the first wave of COVID-19 while the AR process itself was also changing over time. The stock market of the USA followed by Japan performed better than other G7 countries during the first phase of the COVID-19 pandemic while the stock market of France was affected during the COVID-19 pandemic.

Neural Memory State Space Models for Medical Image Segmentation.

With the rapid advancement of deep learning, computer-aided diagnosis and treatment have become crucial in medicine. UNet is a widely used architecture for medical image segmentation, and various methods for improving UNet have been extensively explored. One popular approach is incorporating transformers, though their quadratic computational complexity poses challenges. Recently, State-Space Models (SSMs), exemplified by Mamba, have gained significant attention as a promising alternative due to their linear computational complexity. Another approach, neural memory Ordinary Differential Equations (nmODEs), exhibits similar principles and achieves good results. In this paper, we explore the respective strengths and weaknesses of nmODEs and SSMs and propose a novel architecture, the nmSSM decoder, which combines the advantages of both approaches. This architecture possesses powerful nonlinear representation capabilities while retaining the ability to preserve input and process global information. We construct nmSSM-UNet using the nmSSM decoder and conduct comprehensive experiments on the PH2, ISIC2018, and BU-COCO datasets to validate its effectiveness in medical image segmentation. The results demonstrate the promising application value of nmSSM-UNet. Additionally, we conducted ablation experiments to verify the effectiveness of our proposed improvements on SSMs and nmODEs.

Learning-based Predictive LPV Control of Atmospheric Pressure Plasma Jets

Abstract Complexity of atmospheric pressure plasma jet dynamics poses a significant challenge for control design, and this letter presents a learning- and Scenario-based Model Predictive Control (ScMPC) method in the Linear Parameter-Varying (LPV) framework to tackle this challenge. By leveraging artificial neural networks, an LPV state-space representation of the system dynamics is first learned. The mismatch between this model and real plant is then estimated using Bayesian neural networks, enabling scenario generation for ScMPC design. Soft constraints are imposed in the control design formulation to ensure the feasibility of the underlying optimization problem. Results from extensive simulations are used to compare the proposed framework with a benchmark LTI-based ScMPC, demonstrating superior performance in both reference tracking and thermal dose delivery. The proposed approach allows for accurate control of plasma jets while reducing conservatism inherent in either LTI-based approaches or other robust control methods.

Dynamic optimization algorithms for enhancing blockchain network resilience against distributed attacks

This research introduces a dynamic optimization algorithm designed to enhance blockchain network resilience against distributed attacks such as Distributed Denial of Service (DDoS), Sybil, and eclipse attacks. The primary objective is to develop a real-time, adaptive control strategy that minimizes network performance degradation while dynamically responding to evolving threats. The research design integrates multi-objective optimization, game theory, and reinforcement learning to formulate a defense strategy that adapts to adversarial conditions. The methodology is based on a modified state-space model, where the blockchain's performance is represented by a system of dynamic equations influenced by both control actions (defensive measures) and attack vectors. The optimization problem is formulated to minimize a cost function that balances network resilience and resource usage. A numerical example is presented to validate the model, demonstrating the algorithm’s effectiveness in maintaining network performance under attack by adjusting defense mechanisms in real-time. The main results indicate that the proposed method significantly reduces the impact of distributed attacks while ensuring efficient resource allocation. In conclusion, this research offers a novel framework for enhancing blockchain security, with implications for real-world applications in decentralized systems, financial services, and critical infrastructure. Future work will address the scalability of the algorithm and explore more advanced reinforcement learning techniques to handle more complex and unpredictable attack patterns.

Link Aggregation for Skip Connection–Mamba: Remote Sensing Image Segmentation Network Based on Link Aggregation Mamba

The semantic segmentation of satellite and UAV remote sensing imagery is pivotal for address exploration, change detection, quantitative analysis and urban planning. Recent advancements have seen an influx of segmentation networks utilizing convolutional neural networks and transformers. However, the intricate geographical features and varied land cover boundary interferences in remote sensing imagery still challenge conventional segmentation networks’ spatial representation and long-range dependency capabilities. This paper introduces a novel U-Net-like network for UAV image segmentation. We developed a link aggregation Mamba at the critical skip connection stage of UNetFormer. This approach maps and aggregates multi-scale features from different stages into a unified linear dimension through four Mamba branches containing state-space models (SSMs), ultimately decoupling and fusing these features to restore the contextual relationships in the mask. Moreover, the Mix-Mamba module is incorporated, leveraging a parallel self-attention mechanism with SSMs to merge the advantages of a global receptive field and reduce modeling complexity. This module facilitates nonlinear modeling across different channels and spaces through multipath activation, catering to international and local long-range dependencies. Evaluations on public remote sensing datasets like LovaDA, UAVid and Vaihingen underscore the state-of-the-art performance of our approach.

An ultra-high gain boost converter with low switching stress for integrated multi-energy storage systems

In this paper, a high-gain low-switching-stress coupled-inductor with high voltage step-up voltage multiplier cells quadratic boost converter (VMC-QBC) is proposed. The turn ratio of the coupled inductors and the switch duty cycle increase the dynamic gain, and the two degrees of freedom adjustment and modularity of the voltage multiplier cells (VMC) make the structure more flexible. The use of the same drive signal for both switches makes control easier. While achieving multi-stage boosting and multiplication boosting from low to medium duty cycle, the passive clamping circuit absorbs the energy leaked by the coupled inductor, thus reducing the stress on the switching tube and alleviating the diode reverse recovery problem. A non-ideal model with parasitic parameters is developed to analyse the real voltage gain and the converter losses to give design guidelines. A 300 W prototype is designed and tested. The state space model of the converter is established and the working principle is analysed. Compared to other high-gain quadratic boost converters, the proposed converter has continuous input current, common ground characteristics, and high voltage gain at low to medium duty cycles to accommodate integrated multi-energy storage systems.

Reducing Measurement Costs of Thermal Power: An Advanced MISM (Mamba with Improved SSM Embedding in MLP) Regression Model for Accurate CO2 Emission Accounting.

Current calculation methods for the carbon content as received (Car) of coal rely on multiple instruments, leading to high costs for enterprises. There is a need for a cost-effective model that maintains accuracy in CO2 emission accounting. This study introduces an MISM model using key parameters identified through correlation and ablation analyses. An Improved State-Space Model (ISSM) and an IS-Mamba module are integrated into a Multi-Layer Perceptron (MLP) framework, enhancing information flow and regression accuracy. The MISM model demonstrates superior performance over traditional methods, reducing the Root Mean Square Error (RMSE) by 22.36% compared to MLP, and by 9.65% compared to Mamba. Using only six selected parameters, the MISM model achieves a precision of 0.27% for the discrepancy between the calculated CO2 emissions and the actual measurements. An ablation analysis confirms the importance of certain parameters and the effectiveness of the IS-Mamba module at improving model performance. This paper offers an innovative solution for accurate and cost-effective carbon accounting in the thermal power sector, supporting China's carbon peaking and carbon neutrality goals.

Samba: Semantic segmentation of remotely sensed images with state space model

High-resolution remotely sensed images pose challenges to traditional semantic segmentation networks, such as Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs). CNN-based methods struggle to handle high-resolution images due to their limited receptive field, while ViT-based methods, despite having a global receptive field, face challenges when processing long sequences. Inspired by the Mamba network, which is based on a state space model (SSM) to efficiently capture global semantic information, we propose a semantic segmentation framework for high-resolution remotely sensed imagery, named Samba. Samba utilizes an encoder-decoder architecture, with multiple Samba blocks serving as the encoder to efficiently extract multi-level semantic information, and UperNet functioning as the decoder. We evaluate Samba on the LoveDA, ISPRS Vaihingen, and ISPRS Potsdam datasets using the mIoU and mF1 metrics, and compare it with top-performing CNN-based and ViT-based methods. The results demonstrate that Samba achieves unparalleled performance on commonly used remotely sensed datasets for semantic segmentation. Samba is the first to demonstrate the effectiveness of SSM in segmenting remotely sensed imagery, setting a new performance benchmark for Mamba-based techniques in this domain of semantic segmentation. The source code and baseline implementations are available at https://github.com/zhuqinfeng1999/Samba.

DCSS-UNet: UNet based on State Space Model for Polyp Segmentation

Early and accurate segmentation of medical images can provide valuable information for medical treatment. In recent years, the automatic and accurate segmentation of polyps in colonoscopy images has received extensive attention from the research community of artificial intelligence and computer vision. Many researchers have conducted in-depth research on models based on CNN and Transformer. However, CNN have limited ability to model remote dependencies, which makes it challenging to fully utilize semantic information in images. On the other hand, the complexity of the secondary computation poses a challenge to the transformer. Recently, state-space models (SSMS), such as Mamba, have been recognized as a promising approach. They not only show superior performance in remote interaction, but also maintain linear computational complexity. Inspired by Mamba, we propose DCSS-UNet, where we utilize visual state space (VSS) blocks in VMamba to capture a wide range of contextual information. In the Skip connection phase, we propose Skip Connects Feature Attention modules(SFA) to better communicate information from the encoder. In the decoder stage, we innovatively combined the Temporal Fusion Attention Module(TFAM) to effectively fuse the feature information. In addition, we introduced a custom Loss calculation method, Tversky Loss, for the model to achieve faster convergence and improve segmentation along polyp boundaries. Our model was trained on the Kvasir-SEG and CVC-ClinicDB datasets, and validated on datasets Kvasir-SEG, CVC-ColonDB, CVC-300, and ETIS. The results show that the model achieves good segmentation accuracy and generalization performance with a low number of parameters. We are 6.1% ahead in the Kavirs-SEG dataset and 3.1% ahead in the CVC-ClinicDB dataset compared to VM-UNet.