Distributed Fusion Architecture Research Articles

RADFNet: An infrared and visible image fusion framework based on distributed network.

The fusion of infrared and visible images can improve image quality and eliminate the impact of changes in the agricultural working environment on the information perception of intelligent agricultural systems. In this paper, a distributed fusion architecture for infrared and visible image fusion is proposed, termed RADFNet, based on residual CNN (RDCNN), edge attention, and multiscale channel attention. The RDCNN-based network realizes image fusion through three channels. It employs a distributed fusion framework to make the most of the fusion output of the previous step. Two channels utilize residual modules with multiscale channel attention to extract the features from infrared and visible images, which are used for fusion in the other channel. Afterward, the extracted features and the fusion results from the previous step are fed to the fusion channel, which can reduce the loss in the target information from the infrared image and the texture information from the visible image. To improve the feature learning effect of the module and information quality in the fused image, we design two loss functions, namely, pixel strength with texture loss and structure similarity with texture loss. Extensive experimental results on public datasets demonstrate that our model has superior performance in improving the fusion quality and has achieved comparable results over the state-of-the-art image fusion algorithms in terms of visual effect and quantitative metrics.

Computing Cluster and Intelligent Sensor Network in the Analysis and Application of College Students’ Physical Exercise Behavior

Based on computing cluster and intelligent sensor network technology, in view of network delay, this paper uses first-in-first-out buffers to be built at the node sending and receiving ports to convert the random delay of the physical exercise behavior network control system into a fixed delay. First, we analyze and model the controller design of the physical exercise behavior network control system. Through the analysis and synthesis of the current situation and methods of the physical exercise behavior network control system controller at home and abroad, the sensor is driven by time, and the controller and actuator are used. In the event-driven method, the sending and receiving buffers are set on the network ports of the nodes, the delay is changed from random to fixed at the same time, and the problem of data packet timing disorder is improved. Secondly, through the analysis of the internal control system node, the internal AD, DA conversion, data storage, CPU internal tasks, and task scheduling algorithm modules are implemented in the model. Experimental simulations show that, in view of the difficulty of unsatisfactory tracking effect caused by the aliasing of multiple target signals collected by sensor nodes, a combined tracking strategy is adopted; that is, multiple tracking dynamic clusters are combined into one for tracking when the sports behavior is close. In order to avoid the heavy communication and computing requirements in the centralized mode, mobile sensor networks usually adopt a distributed fusion architecture. The dynamic cluster maintenance and positioning strategy are given. In the stage of separation of multiple sports behaviors, a dynamic cluster decomposition algorithm based on boundary search is proposed, which can effectively determine the degree of separation of sports behaviors and provide a basis for establishing new dynamic clusters for follow-up tracking. The results show that the algorithm can effectively realize the merging and decomposition of dynamic clusters of multiple sports behaviors and effectively realize the dynamic tracking of multiple sports behaviors.

Exponential Mixture Density Based Approximation to Posterior Cramér-Rao Lower Bound for Distributed Target Tracking

The posterior Cramr-Rao lower bound (PCRLB) and a number of its extensions including the unconditional PCRLB (U-PCRLB) and the conditional PCRLB (C-PCRLB) have been widely studied in multi-sensor target tracking (MSTT). Previous MSTT works assume that the measurements are conditionally independent, which is often not the case in practice. When the correlations of the measurements are unknown, the standard Bayes update and PCRLB computations cannot be performed. Inspired by geometric average (GA) fusion that implements data fusion with unknown data correlations by employing exponential mixture density (EMD) to compute the global posterior, EMD-based approximations to U-PCRLB and C-PCRLB are derived in this paper for two classic distributed fusion architectures, namely the hierarchical and consensus architectures. The PCRLB can be decomposed into two parts, one coming from the prior information and the other from the information contained in the data. The data information part of PCRLB is approximated via EMD-based posterior, leading to the approximation of the PCRLB. We present a sequential Monte Carlo solution to recursively compute the proposed approximation to the PCRLB for nonlinear non-Gaussian estimation problems. Numerical simulations are provided to show that the proposed approximation to the bound on the estimation mean square error (MSE) is tighter compared to the existing bounds based on likelihood fusion obtained under the measurement independence assumption.

Distributed fusion in harsh environments using multiple bearings-only sensors with out-of-sequence-refined measurements

This paper investigates a novel distributed fusion system for target tracking in three dimensional surveillance using multiple asynchronous BO sensors by simultaneously addressing the problems of target mis-detection and clutter disturbance as well as the out-of-sequence (OOS) information fusion. The proposed distributed fusion architecture consists of the local tracking and distributed fusion, wherein, the refined BO measurements output from the local tracking in each BO sensor are transmitted to the fusion center for subsequent distributed fusion. In the local tracking, instead of kinematic state, the target pseudo state is defined and tracked by operating the proposed pseudo state-integrated probabilistic data association (PS-IPDA) algorithm with an approximately linear pseudo state evolution model, resulting in greatly refined BO sensor measurements. In the fusion center, the sequential asynchronous IPDA-general one-step retrodiction update (SAIPDA-GORU) algorithm is proposed to effectively fuse the transmitted origin-unknown OOS refined measurements (OOSRMs) with the central tracks. The SAIPDA-GORU propagates the central track hybrid state to the OOS time by deploying a general one-step retrodiction technique, and mathematically formulates the OOSRMs-to-retrodicted central track association probabilities utilized to facilitate the fusing of currently filtered central track with OOSRMs. Numerical results show that the proposed PS-IPDA algorithm reduces the errors of target bearings measurement by around a half and eliminates more than 60% of the clutter measurements contained in raw detections, saving significant communication bandwidth between local sensors and fusion center. Furthermore, the proposed SAIPDA-GORU method tangibly improves the estimation accuracy and false track discrimination over the existing discarding method, with the potential to be implemented in real-time.

A Distributed Fusion Framework of Multispectral and Panchromatic Images Based on Residual Network

Remote sensing images have been widely applied in various industries; nevertheless, the resolution of such images is relatively low. Panchromatic sharpening (pan-sharpening) is a research focus in the image fusion domain of remote sensing. Pan-sharpening is used to generate high-resolution multispectral (HRMS) images making full use of low-resolution multispectral (LRMS) images and panchromatic (PAN) images. Traditional pan-sharpening has the problems of spectral distortion, ringing effect, and low resolution. The convolutional neural network (CNN) is gradually applied to pan-sharpening. Aiming at the aforementioned problems, we propose a distributed fusion framework based on residual CNN (RCNN), namely, RDFNet, which realizes the data fusion of three channels. It can make the most of the spectral information and spatial information of LRMS and PAN images. The proposed fusion network employs a distributed fusion architecture to make the best of the fusion outcome of the previous step in the fusion channel, so that the subsequent fusion acquires much more spectral and spatial information. Moreover, two feature extraction channels are used to extract the features of MS and PAN images respectively, using the residual module, and features of different scales are used for the fusion channel. In this way, spectral distortion and spatial information loss are reduced. Employing data from four different satellites to compare the proposed RDFNet, the results of the experiment show that the proposed RDFNet has superior performance in improving spatial resolution and preserving spectral information, and has good robustness and generalization in improving the fusion quality.

Resource Allocation for Multitarget Tracking and Data Reduction in Radar Network With Sensor Location Uncertainty

Traditional networked radar systems for target tracking usually suffer from a heavy data processing burden, and do not consider the sensor location uncertainties (SLUs), by assuming that radar locations are known perfectly, which is applicable only for static platforms. In this paper, considering sensors mounted on moving platforms, we propose a joint power allocation and measurement selection (JPAMS) strategy for multitarget tracking and data reduction in radar networks with the SLUs. The mechanism is to optimize the transmitted power and select the propagation paths with informative measurements, simultaneously. First, we adopt a distributed fusion architecture to estimate both states of targets and radars in clutter. Based on the distributed fusion architecture, the predicted conditional Cramér-Rao lower bound (PC-CRLB) considering the SLU and the measurement origin uncertainty is derived. Second, the JPAMS strategy is formulated as a bi-objective optimization problem, where the sum of weighted PC-CRLBs and the number of selected propagation paths are used as the performance metrics with respect to tracking and data reduction. The corresponding optimization is a NP-hard problem containing both continuous and binary variables. Third, to solve this nonconvex problem, we propose a sparsity-enhancing sequential convex programming algorithm. Finally, numerical simulations demonstrate the superiority of the proposed JPAMS strategy over the traditional allocation strategies.

An Improved Unscented Particle Filter Approach for Multi-Sensor Fusion Target Tracking.

In this paper, a new approach of multi-sensor fusion algorithm based on the improved unscented particle filter (IUPF) and a new multi-sensor distributed fusion model are proposed. Additionally, we employ a novel multi-target tracking algorithm that combines the joint probabilistic data association (JPDA) algorithm and the IUPF algorithm. To improve the real-time performance of the UPF algorithm for the maneuvering target, minimum skew simplex unscented transform combined with a scaled unscented transform is utilized, which significantly reduces the calculation of UPF sample selection. Moreover, a self-adaptive gain modification coefficient is defined to solve the low accuracy problem caused by the sigma point reduction, and the problem of particle degradation is solved by modifying the weights calculation method. In addition, a new multi-sensor fusion model is proposed, which better integrates radar and infrared sensors. Simulation results show that IUPF effectively improves real-time performance while ensuring the tracking accuracy compared with other algorithms. Besides, compared with the traditional distributed fusion architecture, the proposed new architecture makes better use of the advantages of radar and an infrared sensor and improves the tracking accuracy.

Resource Scheduling for Distributed Multi-Target Tracking in Netted Colocated MIMO Radar Systems

In this paper, an effective solution is proposed for joint beam and power scheduling (JBPS) in the netted Colocated MIMO (C-MIMO) radar systems for distributed multi-target tracking (MTT). At its core, the proposed solution includes a distributed fusion architecture that reduces the communication requirements while maintaining the overall robustness of the system. The distributed fusion architecture employs the covariance intersection (CI) fusion to address the unknown information correlations among radar nodes. Each C-MIMO radar node in the network can generate a time-varying number of beams with controllable transmitting power by waveform synthesis, thus is capable of accomplishing multiple tracking tasks simultaneously. To maximize the global MTT performance of the radar network, the proposed JBPS solution implements an online resource scheduling, regarding both the generated beams and the transmitted power of all radar nodes, based on the feedback of the MTT results. A scaled accuracy-based objective function is designed to quantify the global MTT performance while properly taking into account different target priorities on resource allocation. The Bayesian Cramer-Rao lower bound (BCRLB) for CI fusion rule is derived and utilized as the constituent of the objective function since it provides a lower bound on the accuracy of the target state estimates. As the formulated JBPS problem is non-convex, we propose a fast reward-based iterative descending approach to solve it effectively. Numerical results show that the proposed JBPS can deliver superior performance in terms of maximizing the overall MTT performance while possessing high flexibility on the resource allocation regarding different target priorities.

Target tracking based on improved square root cubature particle filter via underwater wireless sensor networks

Target tracking in underwater wireless sensor networks (UWSNs) has two fundamental issues that tracking accuracy and energy consumption. Although the application of particle filter in target tracking is considered in recent years, degeneracy phenomenon and particle impoverishment always restrict its capacity and application. This paper proposes an improved square root cubature particle filter (ISRCPF) to improve tracking accuracy. The authors employ self-adaptive artificial fish swarm algorithm (AFSA) to optimise the particles, which makes the particles move towards to high likelihood region and maintains the diversity of the particles. Moreover, a sensor selection scheme is provided, which reduces energy consumption of the network by exactly waking up four sensor nodes at each time, while preserving tracking accuracy. Additionally, the authors propose a novel fusion method called similarity fusion method (SFM) to fuse local estimates together and then obtain a better result for distributed fusion architecture (DFA). The simulation results demonstrate that the proposed methods have superior performance.

Target tracking based on improved cubature particle filter in UWSNs

In this study, an improved cubature particle filter based on the artificial bee colony (ABC) algorithm is proposed and applied to target tracking via underwater wireless sensor networks (UWSNs). In the proposed method, the square root cubature Kalman filter is used to generate the proposal distribution and the ABC algorithm is employed to optimise the particles before resampling, which makes the particles move toward the high likelihood region and maintain the diversity of the particles. Moreover, linear minimum variance criterion is utilised to fuse local estimates together in distributed fusion architectures of UWSNs. The simulation results show that the proposed method outperforms other classical algorithms in tracking accuracy.

Multi‐sensor track‐to‐track fusion with target existence in cluttered environments

Multi-sensor fusion for multiple target tracking in cluttered environments is needed for improving tracking accuracy and track maintenance over single sensor target tracking in real target tracking applications. A track quality measure, such as the probability of target existence, can be used when fusing local sensor tracks for false track discrimination at the fusion centre and to improve the performance of the track-to-track association. The track quality measure is a highly effective tool for track maintenance and lowers the computational load by sending tracks with a high probability of target existence to the track centre. In this study, the authors utilise the iterative joint-integrated probabilistic data association technique for multi-sensor distributed fusion systems. Track-to-track fusion performance also depends on distributed fusion architecture. The track-to-track fusion algorithm is evaluated in distributed fusion systems without memory, with memory, and with feedback. The track confirmation rates and position errors are simulated and compared with low-level centralised fusion, which is theoretically optimal.

분산형 다중 레이다 융합시스템을 위한 추적 관리 구조 선정

To obtain accurate target information, fusion technique can be used in a multiple radar air defense system. This article proposes a stable track management architecture for hierarchical distributed multiple radar fusion system which consists of multiple subordinate radars and superposed fusion center. It is difficult to apply Track-to-Track association and Track-to-Track fusion technique, which are a basic component of distributed fusion architecture, directly to a radar system which requires high reliability. This paper selected fusion architecture and Track-to-Track fusion method which can improve track fusion performance, and suggested stable track management architecture which contains track merge component.

A novel distributed fusion algorithm for multi-sensor nonlinear tracking

The covariance intersection (CI), especially with feedback structure, can be easily combined with nonlinear filters to solve the distributed fusion problem of multi-sensor nonlinear tracking. However, this paper proves that the CI algorithm is sub-optimal, thus degrading the fusion accuracy. To avoid such an issue, a novel distributed fusion algorithm, namely Monte Carlo Bayesian (MCB) algorithm, is proposed. First, it builds a distributed fusion architecture based on the Bayesian tracking framework. Then, the Monte Carlo sampling is incorporated into this architecture to form a feasible solution to nonlinear tracking. Finally, the simulation results verify that our MCB algorithm advances the state-of-the-art distributed fusion of nonlinear tracking.

Dual-Module Data Fusion of Infrared and Radar for Track Before Detect

A track before detect method based on data fusion of infrared and radar is proposed to increase the probability of correct track initiation and shorten initiation time. Track before detect is a new technique for dim target detection and tracking which is useful when the signal-to-noise ratio of target is low. Particle filter and dynamic programming for track before detect are currently proposed for detecting and tracking dim targets in low signal-to-noise ratio background. In this paper, we apply them in dual-module data fusion of infrared and radar. Particle Filter is applied to process the acquired data from radar. Dynamic programming is applied to process the acquired data from Infrared. Sensor receives data and generates the first stage decision. The decisions are subsequently transmitted to the fusion center where they are combined into a final decision on distributed fusion architectures. The proposed method is applied to simulate track before detect in dual module system of infrared and radar. The simulation results show that the proposed method increase the probability of correct track initiation and shorten initiation time.

Triangle orientation discrimination performance model for a multiband IR imaging system with human vision

In support of multiband imaging system performance forecasting, an equation-based triangle orientation discrimination (TOD) model is developed. Specifically, with the characteristic of the test pattern related to spectrum, the mathematical equations for predicting the TOD threshold of the system with distributed fusion architecture in the IR spectrum band are derived based on human vision with the "k/N" fusion rule, with emphasis on the impacts of fusion on the threshold. Furthermore, a figure of merit Q related to the TOD calculation results is introduced to analyze the relation of the discrimination performance of multiband imaging system to the size and the spectral difference of test pattern. The preliminary validation with the experiment results suggests that our proposed model can provide a reasonable prediction of the performance for a multiband imaging system.

Distributed Fusion Architectures and Algorithms for Target Tracking

Modern surveillance systems often utilize multiple physically distributed sensors of different types to provide complementary and overlapping coverage on targets. In order to generate target tracks and estimates, the sensor data need to be fused. While a centralized processing approach is theoretically optimal, there are significant advantages in distributing the fusion operations over multiple processing nodes. This paper discusses architectures for distributed fusion, whereby each node processes the data from its own set of sensors and communicates with other nodes to improve on the estimates, The information graph is introduced as a way of modeling information flow in distributed fusion systems and for developing algorithms. Fusion for target tracking involves two main operations: estimation and association. Distributed estimation algorithms based on the information graph are presented for arbitrary fusion architectures and related to linear and nonlinear distributed estimation results. The distributed data association problem is discussed in terms of track-to-track association likelihoods. Distributed versions of two popular tracking approaches (joint probabilistic data association and multiple hypothesis tracking) are then presented, and examples of applications are given.