Low Weight Particles Research Articles

Intelligent enhanced particle filter with deep residual network surrogate for accurate groundwater pollution source characterization

Accurate groundwater pollution source characterization (GPSC) is crucial for environmental protection and water resource management. Although the particle filter (PF) algorithm is widely employed for GPSC, it suffers from particle degradation and loss of diversity. To address this challenge, this study proposes an intelligent enhanced particle filter (IEPF) method to enhance GPSC’s accuracy and efficiency. First, this method optimizes the resampling process by defining a weight threshold while introducing an information transmission mechanism. This mechanism enables low-weight particles to update their positions by leveraging information from high-weight particles and neighboring particles, thereby enhancing information exchange among particles and improving the algorithm’s global search capability. Crossover and mutation operations are introduced to further increase the particle diversity, effectively preventing particles from concentrating on local optimal solutions. Furthermore, this study employs a deep residual network (ResNet) as a surrogate model for the simulation model. The ResNet has strong feature learning and expression capabilities and efficiently captures nonlinear relationships in data, thus significantly improving computational efficiency while ensuring accuracy. The results demonstrate that the IEPF method significantly enhances GPSC accuracy. In summary, the proposed IEPF method combined with the ResNet surrogate model provides an efficient and accurate approach to GPSC, with considerable benefits for groundwater pollution management.

An Intelligent Cost-Reference Particle Filter with Resampling of Multi-Population Cooperation.

Although the cost-reference particle filter (CRPF) has a good advantage in solving the state estimation problem with unknown noise statistical characteristics, its estimation accuracy is still affected by the lack of particle diversity and sensitivity to the particles' initial value. In order to solve these problems of the CRPF, this paper proposed an intelligent cost-reference particle filter algorithm based on multi-population cooperation. A multi-population cooperative resampling strategy based on ring structure was designed. The particles were divided into multiple independent populations upon initialization, and each population generated particles with a different initial distribution. The particles in each population were divided into three different particle sets with high, medium and low weights by the golden section ratio according to the weight. The particle sets with high and medium weights were retained. Then, a cooperative strategy based on Gaussian mutation was designed to resample the low-weight particle set of each population. The high-weight particles of the previous population in the ring structure were randomly selected for Gaussian mutation to replace the low-weight particles in the current population. The low-weight particles of all populations were resampled in turn. The simulation results show that the intelligent CRPF based on multi-population cooperation proposed in this paper can reduce the sensitivity of the CRPF to the particles' initial value and improve the particle diversity in resampling. Compared with the general CRPF and intelligent CRPF with adaptive MH resampling (MH-CRPF), the RMSE and MAE of the proposed method are lower.

Adaptive genetic algorithm-based particle herding scheme for mitigating particle impoverishment

Particle filters (PFs) estimate the true parameter state from samples of states, called particles, that are drawn to construct an approximated posterior probability density function (PDF). Each drawn particle has its own probability of being the true state called importance weight. Both particle degeneracy and impoverishment are important problems found in generic PFs that cause particles to become trapped at local maxima. Genetic algorithms (GAs) were applied to herd the drawn particles to discover new high-likelihood state values in order to reshape the approximated posterior PDF and to enhance the accuracy of state estimation, while the particle diversity is maintained. We calculated the threshold weight for classifying particles as low-weight particles and high-weight particles. Each parent pair consists of a high-weight particle and a low-weight particle. High-weight particles are always kept unchanged. After each new offspring particle is found using GA operations, we must check its weight to ensure that it is actually superior to its low-weight parent. Otherwise, this offspring must not be allowed to replace its low-weight parent and must be destroyed. To ensure that the number of high-weight particles is adequate, some offspring particles evolve to be the new high-weight parent particles to produce high-weight offspring with further promoted diversity. Results showed that our approach significantly improves posterior PDF approximation and reduces errors in state estimation.

Research on SLAM Road Sign Observation Based on Particle Filter.

With the development of computer hardware technology, the real-time problem of visual target tracking algorithm increasingly depends on hardware solutions. The core problem of visual target tracking is how to enhance the robustness of tracking algorithm to various complex background environments and various interference factors. Aiming at overcoming the defect that the traditional SLAM (simultaneous localization and map building) algorithm based on EKF (extended Kalman filter) has a slow repair speed for environmental interference, a Monocular SLAM_WOCPF (Monocular vision SLAM based on weight optimization combined particle filter) algorithm is proposed. The weights of all particles are reoptimized in the particle set and they are combined with the tendency of particles to degenerate and deplete. In this way, the chance of self replication of low weight particles is increased, thus increasing the diversity of the whole sample. Furthermore, the improved PF (particle filter) algorithm is applied to solve the problem of road sign observation of mobile robots, so as to expand its application scope. The results show that the mean road sign errors of the Monocular SLAM_WOCPF algorithm in two noise environments are 0.332/m and 0.441/m. The conclusion shows that the Monocular SLAM_WOCPF road sign observation method proposed in this paper can effectively improve the matching success rate of visual road signs and improve the observation quality.

Design of robust cubature fission particle filter algorithm in multi-source cooperative navigation

As a part of the multi-source cooperative navigation scheme, data fusion has a significant impact on the quality of state estimation. Particle filtering has gradually become the focus of many fusion methods due to its unique theoretical advantages in nonlinear non-Gaussian systems. However, the particle degradation and the resulting sample impoverishment restrict its application in complex engineering scenarios. In this paper, a robust cubature fission particle filter (RCFPF) is proposed to deal with these problems. First, in the framework of cubature rule, Huber function is used to combine the L2 norm and L1 norm to improve the importance density function(IDF), suppress the observation noise. Meanwhile, the proposed distribution(PD) is further optimized by combining the Gaussian distribution with Laplace distribution to alleviate particle degradation. Second, the particle swarm is fissioned before resampling, and the particle weight is reconstructed by fission of high weight particles and covering low weight particles to inhibit sample impoverishment. The vehicle experiments of multi-source cooperative navigation show that the proposed algorithm achieves better test results in accuracy and robustness than extended Kalman filter (EKF), strong tracking particle filter (STPF), and cubature particle filter (CPF).

Improved Grid-Based Rao-Blackwellized Particle Filter SLAM Based on Grey Wolf Optimizer

In this work, a Rao-Blackwellized particle filter simultaneous localization and mapping based on grey wolf optimizer (called GWO-RBPF) is proposed. The proposed method aims to improve the accuracy of the mapping while maintaining the number of particles. GWO-RBPF utilizes the local exploration and global development ability of the grey wolf optimizer to improve the estimation performance of the Rao-Blackwellized particle filter, so that the low-weight particles can approach high-weight particles. Meanwhile, the pose information of the particles is optimized by the grey wolf optimizer. The proposed method is applied to the benchmark datasets and real-world datasets. The experimental results show that our method outperforms conventional method in terms of map accuracy versus the number of particles.

Moving Target Localization in Sports Image Sequence Based on Optimized Particle Filter Hybrid Tracking Algorithm

This paper proposes a fusion stadium positioning algorithm, which uses self-optimizing particle filter to integrate the improved athlete dead reckoning and WiFi position fingerprint algorithm for stadium positioning. In order to determine the initial absolute position of the stadium positioning, for athletes entering the stadium from the outside, a seamless switching algorithm outside the stadium is proposed, using the characteristics of high-altitude satellite GPS to find a suitable switching point as the initial absolute position. If in the stadium, WiFi static positioning determines the initial absolute position. Then, aiming at the problem that the poorly diversified particles cannot be better integrated and localized, a self-optimized particle filter algorithm is proposed. After resampling and retaining high-weight particles, the characteristics of low-weight particles are embedded in the copied high-weight particles. This can improve diversity, and we finally carry out fusion positioning. The target tracking algorithm based on Mean Shift has a fixed-scale tracking window, and the tracking effect of variable-size targets is not ideal. In this paper, an affine transformation algorithm is introduced to improve it. First, we iterate the adjacent image frames in reverse Mean Shift to determine the center position of the target and then use the corner matching method to perform template matching on the target to adjust the size of the tracking window. Through simulation verification, it is proved that the optimized particle filter hybrid tracking algorithm can achieve the ideal result when the target size changes. For the image sequence S1, the tracking window of the 20th frame and the 40th frame has a small offset, but the optimal position can be quickly found by Mean Shift iteration. For the image sequence S2, between the 40th frame and the 60th frame, the target occlusion causes the accuracy of the target template to decrease, and the Bhattacharyya coefficient is at a relatively low value. For the image sequence S3, the tracking effect of the optimized particle filter hybrid tracking algorithm meets the requirements.

Laser SLAM optimization based on resampling

In order to solve the particle degradation and particle exhaustion in the simultaneous positioning and map construction method of Rao-Blackwellized particle filter, an optimization method for simultaneous positioning and map construction is proposed. In order to alleviate the need for too much data in the re-sampling stage, the calculation time is longer, and the calculation amount is reduced by dropping some low-weight particles and reducing the number of particles. The experimental results show that compared with the traditional RBPF-SLAM method, the improved method can obtain a more accurate map with fewer particles, reduce the number of Kalman particle filtering, and reduce the amount of calculation to meet the needs of actual use.

An Intelligent Particle Filter With Adaptive M-H Resampling for Liquid-Level Estimation During Silicon Crystal Growth

During the growth of silicon single crystals, it is critical to detect the liquid level of the silicon melt to ensure their high-quality production. Because noise statistics are difficult to determine in measured values of the liquid level, a particle filter (PF) with unknown statistics has been presented to estimate the liquid level. However, this approach leads to inaccurate results due to sample impoverishment. To alleviate this problem, we propose an intelligent PF method with an adaptive Metropolis–Hastings (M-H) resampling strategy. To accomplish this, we first design an M-H resampling strategy with two proposed distributions to resample low-weight particles. These distributions randomly select high-weight particles for the Gaussian mutations or high-weight and low-weight particles for crossover operations, so as to promote the movement of low-weight particles to high-probability regions. We also construct a self-adaptive function to further improve the overall particle quality, which is used to calculate the selection probability of these two proposed distributions according to the proportion of low-weight particles in all of the particles. Finally, the liquid level is estimated according to the particles after the modified resampling strategy is applied. A comparative evaluation of the proposed method with the adaptive genetic particle filter (AGPF) and the firefly algorithm intelligence optimized particle filter (FAIOPF) is conducted. Some results of the simulation and the practical experiment are presented; they indicate the proposed method offers accuracy improvements in the liquid-level estimation during the silicon crystal growth. More specifically, compared with the AGPF and the FAIOPF, the mean absolute error (MAE) of the proposed method has been reduced by approximately 53.3% and 99.5%, respectively.

Pose Selection and Feedback Methods in Tandem Combinations of Particle Filters with Scan-Matching for 2D Mobile Robot Localisation

Robot localisation is predominantly resolved via parametric or non-parametric probabilistic methods. The particle filter, the most common non-parametric approach, is a Monte Carlo Localisation (MCL) method that is extensively used in robot localisation, as it can represent arbitrary probabilistic distributions, in contrast to Kalman filters, which is the standard parametric representation. In particle filters, a weight is internally assigned to each particle, and this weight serves as an indicator of a particle’s estimation certainty. Their output, the tracked object’s pose estimate, is implicitly assumed to be the weighted average pose of all particles; however, we argue that disregarding low-weight particles from this averaging process may yield an increase in accuracy. Furthermore, we argue that scan-matching, treated as a prosthesis of (or, put differently, fit in tandem with) a particle filter, can also lead to better accuracy. Moreover, we study the effect of feeding back this improved estimate to MCL, and introduce a feedback method that outperforms current state-of-the-art feedback approaches in accuracy and robustness, while alleviating their drawbacks. In the process of formulating these hypotheses we construct a localisation pipeline that admits configurations that are a superset of state-of-the-art configurations of tandem combinations of particle filters with scan-matching. The above hypotheses are tested in two simulated environments and results support our argumentation.

An Adaptive Particle Filter Technique for System State Estimation and Prognosis

System state estimation and prognostics are the key issues in dynamic system monitoring and management. Although the particle filter (PF) has been applied to model the nonlinear degradation feature of the system aging mechanism in several studies, it has two potential problems: the sample degeneracy and the impoverishment. To tackle these problems, a new adaptive particle filter (APF) technique is proposed in this article to enhance the performance of PFs. In the APF, a self-evaluation method is suggested to track the posterior distribution and detect the low-weight particles (sample degeneracy). A new adaptive weight adjustment approach is proposed to adaptively explore the posterior space, process those low-weight particles, and tackle the sample degeneracy. The effectiveness of the proposed APF technique is validated by simulation tests using several model conditions. It is also implemented for battery-health monitoring and prognosis. Test results show that the proposed APF technology can effectively capture and track the system’s dynamic characteristics.

An Enhanced Mutated Particle Filter Technique for System State Estimation and Battery Life Prediction

The particle filter (PF) has been used for the analysis of nonlinear, non-Gaussian dynamical systems with hidden state variables. However, PF has some limitations in real-world applications, for instant, the sample degeneracy and the impoverishment, which are considered as long-standing challenges in this research and development field. Although several techniques have been proposed in the literature for this purpose, they have some limitations: for example, they cannot represent the entire probability density function (pdf) effectively and are usually problem dependent. In this paper, an enhanced mutated PF (EMPF) technique is proposed to improve the performance of PFs. In the EMPF technique, first, a novel enhanced mutation approach is proposed to actively explore the posterior pdf to locate the high-likelihood area. Second, a new selection scheme is suggested to process low-weight particles for optimizing the posterior distribution and tackling sample degeneracy. Third, an outlier assessment method is adopted to monitor the overall pattern of the posterior distribution based on the interquartile range statistical analysis. The effectiveness of the proposed EMPF technique is verified by simulation tests. It is also implemented for the remaining useful life prediction of lithium-ion batteries. Test results show that the developed EMPF technique can capture a system’s dynamic behavior and track system characteristics effectively.

Predicting the travel time of arterial traffic using particle filter with speed matrix

Travel time prediction is an essential part of intelligent transportation system applications. However, the existing travel time prediction methods mainly focus on the freeway due to its simplicity and the high coverage of sensors and few researches have been conducted for the urban arterial road. Consequently, a travel time prediction algorithm based on particle filter is proposed in this paper to predict short-term travel time of the arterial traffic with historical floating car data and the concept of speed matrix is developed to illustrate the spatiotemporal properties of the arterial traffic. Unlike previous travel time prediction methods, the proposed algorithm uses particles with corresponding weights to model the traffic trend in the historical data instead of state-transition function and the weight for each particle are calculated with similarities between the speed matrix of the particle and the current traffic pattern. Moreover, a resampling process is developed to solve the degeneracy problem of the particles by replacing the low-weight particles with historical data. A real floating car dataset of 10357 taxis over a period of 3 months within Beijing is utilized to evaluate the performances of the algorithms. The proposed algorithm has the least errors by comparing with other three algorithms.

Emergency-Oriented Spatiotemporal Trajectory Pattern Recognition by Intelligent Sensor Devices

This paper presents an emergency-oriented procedure to recognize trajectory patterns by analyzing GPS data collected from intelligent sensor devices. An overall description, including design architecture and system modules, is presented. The primary issues are devoted to satisfying the requirements of key group identification and surveillance under normal and emergency circumstance. For the sake of panoramic understanding of human distribution and movement, semantic trajectory information is extracted from dynamic transportation data and static human distribution data. The sequential Monte Carlo method in conjunction with a state-transition model is employed to predict the updating real-time locations. The proposed algorithm selects particles from time-stamped sequential historical data sets. Simultaneously, a resampling strategy is developed to replace low-weight particles. A curve similarity measurement called Fréchet distance is employed to compare trajectories and city roads. Afterward, human daily location and significant locations are identified based on the clustering method. To evaluate the proposed procedure and methods, sequential trajectory data sets come from the GeoLife project along with human distribution logs from smartphone application EMAPP are utilized. Finally, we demonstrate the potential of dealing location information for promoting emergency management.

GPS Receiver Autonomous Integrity Monitoring Algorithm Based on Improved Particle Filter

Reliability of a navigation system is one of great importance for navigation purposes. Therefore, an integrity monitoring system is an inseparable part of aviation navigation system. Failures or faults due to malfunctions in the systems should be detected and repaired to keep the integrity of the system intact. According to the characteristic of GPS (Global Positioning System) receiver noise distribution and particle degeneracy and sample impoverishment problem in particle filter, an improved particle filter algorithm based on genetic algorithm for detecting satellite failures is proposed. The combination of the re-sampling method based on genetic algorithm and basic particle filter is used for GPS receiver autonomous integrity monitoring (RAIM). Dealing with the low weight particles on the basis of the genetic operation, genetic algorithm is used to classify the particles. It brings the selection, crossover and mutation operation in genetic algorithm into the basis particle filter .The method for detecting satellite failures which affect only subsets of system measurement. In addition to a main particle filter, which processes all the measurements to give the optimal state estimate, a bank of auxiliary particle filters is also used, which process subsets of the measurements to provide the state estimates which serve as failure detection references. The consistency of test statistics for detection and isolation of satellite fault is established. The failure detection is undertaken by checking the system state logarithmic likelihood ratio (LLR). The RAIM algorithm combined the genetic particle filter and the likelihood method is illustrated in detail. Experimental results based on GPS real raw data demonstrate that the algorithm under the condition of non- Gaussian measurement noise can improve the accuracy of state estimation, effectively detect and isolate fault satellite, improve the performance of the fault detection. Experimental results demonstrate that the proposed approach is available and effective for GPS RAIM.