Improve Localization Performance Research Articles

An Enhanced Direct Position Determination of Mixed Circular and Non-Circular Sources Using Moving Virtual Interpolation Array.

In this study, a moving single-station direct position determination (DPD) algorithm based on virtual interpolated arrays is proposed. Existing moving single-station algorithms face challenges such as the incomplete utilization of sparse array apertures and insufficient consideration of mixed circular and non-circular signals. To address these issues, we propose an enhanced gridless DPD algorithm, suitable for multiple mixed circular and non-circular sources. Through constructing a non-zero unconjugated covariance matrix from the non-circular components of the mixed signals, the data dimensionality is expanded, and the gridless method is used to fill the voids in the coarray, significantly improving localization performance. Additionally, a unitary transformation method is applied to reduce computational complexity. This method transforms complex operations into real operations by applying unitary transformations to steering vectors and subspaces. Simulation results demonstrate that the proposed algorithm offers significant advantages in terms of array degrees of freedom and localization accuracy.

Carrier Phase-Based Underwater Source Localization for Ultrashort Baseline

In this paper, we propose a phase-based pseudorange estimation and source localization algorithm to improve the localization performances of ultrashort baseline (USBL) systems. The conventional USBL system estimates pseudorange through cross-correlation of received chirps. However, due to the limited sampling rate, precise pseudorange cannot be estimated, resulting in low localization performance. The proposed method calculates the number of wavelengths between the transducer and receiver and detects the initial phase to estimate the pseudorange more accurately than the conventional method. It also improves localization performance by estimating and compensating for phase variations by underwater channels. The source localization performances of the proposed method and the conventional methods were compared and analyzed through computer simulations and ocean experiments, and the proposed method had better localization performances than the conventional methods.

Image manipulation localization via dynamic cross-modality fusion and progressive integration

Image manipulation localization aims at distinguishing forged regions from the whole test image. Though many outstanding prior arts have been proposed for this task, there are still two issues that need to be further studied: (1) how to fuse diverse types of features with forgery clues; (2) how to progressively integrate multistage features for better localization performance. In this paper, we propose a tripartite progressive integration network (TriPINet) for end-to-end image manipulation localization. First, we extract both visual perception information, e.g., RGB input images, and visual imperceptible features, e.g., frequency and noise traces for forensic feature learning. Second, we develop a guided cross-modality dual-attention (gCMDA) module to fuse different types of forged clues. Third, we design a set of progressive integration squeeze-and-excitation (PI-SE) modules to improve localization performance by appropriately incorporating multiscale features in the decoder. Extensive experiments are conducted to compare our method with state-of-the-art image forensics approaches. The proposed TriPINet obtains competitive results on several benchmark datasets.

Enhancing fruit and vegetable detection in unconstrained environment with a novel dataset

Automating the detection of fruits and vegetables using computer vision is essential for modernizing agriculture, improving efficiency, ensuring food quality, and contributing to sustainable and technologically advanced farming practices. This paper presents an end-to-end pipeline for detecting and localizing fruits and vegetables in real-world scenarios. To achieve this, a dataset named FRUVEG67 was curated that includes images of 67 classes of fruits and vegetables captured in unconstrained scenarios, with only a few manually annotated samples per class. A semi-supervised data annotation algorithm (SSDA) was developed that generates bounding boxes for objects to label the remaining nonannotated images. For detection, Fruit and Vegetable Detection Network (FVDNet) was proposed, an ensemble version of YOLOv8n featuring three distinct grid configurations. In addition, an averaging approach for the prediction of the bounding box and a voting mechanism for the prediction of the classes was implemented. Finally Jensen–Shannon Divergence (JSD) in conjunction with focal loss was integrated as the overall loss function for better detection of smaller objects. Experimental results highlight the superiority of FVDNet compared to recent versions of YOLO, showcasing remarkable improvements in detection and localization performance. An impressive mean average precision (mAP) score of 0.82 across all classes was achieved. Furthermore, the efficacy of FVDNet on open-category refrigerator images were evaluated, where it demonstrates promising results.

Auditory Localization Performance in Cochlear Implant Recipients With Single-Sided Deafness: The Challenges and Limitations of Current Outcome Metrics.

Acoustic localization accuracy metrics currently employed in clinical literature both overestimate and underestimate performance benefit of cochlear implantation (CI) for single-sided deafness (SSD). Although localization in SSD with CI has been investigated, performance characterization has relied heavily on average error. Although attractively concise, this measure may misrepresent performance. Here, we characterize frequency-specific localization on a granular level in subjects with CI for SSD as a critical analysis of localization outcome metrics. Eight CI recipients with SSD were recruited. Stimuli of broadband (BBN) and narrowband noise (NBN) at low (500 Hz), mid (1000 Hz), and high (4000 Hz) frequencies were presented in a semianechoic chamber. Localization accuracy was quantified in mean angular error (MAE) and linear regression slope. Use of a CI for SSD subjects improved localization performance by slope for all stimuli ( p ≤ 0.0033) to a level that was equal to normal-hearing controls at 1 and 4 kHz ( p ≥ 0.2281). MAE was also significantly improved for SSD subjects using CI for BBN stimuli ( p ≪ 0.0001); however, no statistically significant improvement in MAE was seen for NBN ( p ≥ 0.5773) with CI use. Descriptive analysis of individual subject performance highlights the reasons for contradictory results. There is inherent challenge in characterizing localization benefit for individuals with CI for SSD. Our data demonstrate the limitations of utilization of average error as the sole metric for outcome benefit. We emphasize the importance of continued research investigating alternative outcome measures as we work toward a more refined understanding of the potential benefits and limitations of cochlear implantation for SSD.

Enhancing Single-Frame Supervision for Better Temporal Action Localization.

Temporal action localization aims to identify the boundaries and categories of actions in videos, such as scoring a goal in a football match. Single-frame supervision has emerged as a labor-efficient way to train action localizers as it requires only one annotated frame per action. However, it often suffers from poor performance due to the lack of precise boundary annotations. To address this issue, we propose a visual analysis method that aligns similar actions and then propagates a few user-provided annotations (e.g., boundaries, category labels) to similar actions via the generated alignments. Our method models the alignment between actions as a heaviest path problem and the annotation propagation as a quadratic optimization problem. As the automatically generated alignments may not accurately match the associated actions and could produce inaccurate localization results, we develop a storyline visualization to explain the localization results of actions and their alignments. This visualization facilitates users in correcting wrong localization results and misalignments. The corrections are then used to improve the localization results of other actions. The effectiveness of our method in improving localization performance is demonstrated through quantitative evaluation and a case study.

Subarray-based joint source localization in shallow water waveguide via subspace intersection

By the use of an underwater bottom-mounted Horizontal Linear Array (HLA), the two-dimensional joint localization technique and its accuracy are analyzed in this article. The proposed method takes advantage of arrival angles measured by a dual-subarray system consisting of a large-aperture HLA. Then, for the system, expressions for source position estimate and error variation are derived. Analysis of the Geometrical Dilution of Precision, which is a measure of localization error defined in Eq. (4), reveals that the localization error is mainly related to the source bearing, subarray interval and their measurement errors. In order to investigate the impact of the underwater waveguide, numerical simulations were carried out based on the normal mode theory. By using the distribution of acoustic intensity and structure of modal arrival, it is proved that the localization accuracy increases as the grazing angle and the subarray interval increase. However, when the measurement errors of source bearing and subarray interval increase, the localization accuracy will decrease. In addition, the obtained bearing estimation result provides an evident characteristic of beam spreading. When sources are near the end-fire direction of the dual-subarray system, significant bearing estimation error occurs, the negative effect of which is then presented. By applying the Subspace Intersection method, unbiased bearing estimation has been achieved, with relatively reduced computational complexity. Most of all, the improved localization performance is validated by experimental data collected in the South China Sea.

EC-Net: General image tampering localization network based on edge distribution guidance and contrastive learning

General Image Tampering Localization (GITL) is a challenging task that aims to locate tampered regions in manipulated images accurately. Despite notable achievements in recent years, existing methods still struggle with generalization and pixel-level segmentation, particularly in the face of diverse tampering forms and invisible traces. To address these issues, we propose EC-Net, a GITL network based on edge distribution guidance and contrastive learning. EC-Net decomposes the task into two phases: the coarse locating phase and the fine locating phase. In the coarse locating phase, a contrastive learning strategy is incorporated to improve the model's generalization capabilities. Additionally, an edge distribution prediction method is introduced to anticipate the global contour information of tampered regions. In the fine locating phase, we introduced a cross-scale fusion mechanism to synthesize feature information across different scales, enhancing the model's ability to discern subtle changes in tampered regions. Additionally, an edge reconstruction mechanism is introduced to restore details, thereby improving the accuracy of the final localization results. Evaluation on seven datasets shows that EC-Net significantly outperforms seven state-of-the-art methods, improving localization performance by 0.095 (F1), 0.088 (IoU), and 0.043 (AUC) compared to the second-best method on average. Additionally, we present a tampered image dataset (RLS26K) for model auxiliary training, featuring diverse real-life scenes with various tampering types, diverse styles, and rich content.

BLE-Based Indoor Localization: Analysis of Some Solutions for Performance Improvement.

This paper addresses indoor localization using an anchor-based system based on Bluetooth Low Energy (BLE) 5.0 technology, adopting the Received Signal Strength Indicator (RSSI) for the distance estimation. Different solutions have been proposed in the scientific literature to improve the performance of this localization technology, but a detailed performance comparison of these solutions is still missing. The aim of this work is to make an experimental analysis combining different solutions for the performance improvement of BLE-based indoor localization, identifying the most effective one. The considered solutions involve different RSSI signals' conditioning, the use of anchor-tag distance estimation techniques, as well as approaches for estimating the unknown tag position. An experimental campaign was executed in a complex indoor environment, characterized by the continuous presence in the movement of working staff and numerous obstacles. The exploitation of multichannel transmission using RSSI signal aggregation techniques showed the greater performance improvement of the localization system, reducing the positioning error (from 1.5 m to about 1 m). The other examined solutions have shown a lesser impact in the performance improvement with a decrease or an increase in the positioning errors, depending on the considered combination of the adopted solutions.

Sparse Bayesian learning based on spatio-temporal structure-aware for matched field processing.

In the application of the matched field processing (MFP) algorithm for underwater acoustic source localization, the measurements at each time step are conventionally processed independently. This study incorporates the prior information about the continuous spatial changes of the source over time under realistic conditions, a factor anticipated to improve localization performance. In this paper, a sparse Bayesian learning (SBL) algorithm based on the spatio-temporal structure-aware is described. We exploit a structure prior for sparse coefficients to capture the continuous spatial structure between adjacent time steps. Moreover, the sparse coefficient can automatically select the update method, utilizing the statistical information from adjacent neighbors or updating independently. The hidden variables in the hierarchical Bayesian framework are inferred via variational Bayesian inference (VBI). Additionally, we extend the proposed method to the multi-frequency case. This method inherits the advantages of the SBL and further reduces position estimation errors. Compared to other approaches, the construction of an accurate motion model is not required. The efficacy of the proposed algorithm is demonstrated through simulation examples and an analysis of the SWellEx-96 experimental data.

Development and Human Factors Evaluation of a Portable Auditory Localization Training System.

To design and develop a Portable Auditory Localization Acclimation Training (PALAT) system capable of producing psychoacoustically accurate localization cues; evaluate the training effect against a proven full-scale, laboratory-grade system under three listening conditions; and determine if the PALAT system is sensitive to differences among electronic level-dependent hearing protection devices (HPDs). In-laboratory auditory localization training has demonstrated the ability to improve localization performance with the open (natural) ear, that is, unoccluded, and while wearing HPDs. The military requires a portable system capable of imparting similar training benefits as those demonstrated in laboratory experiments. In a full-factorial repeated measures design experiment, 12 audiometrically normal participants completed localization training and testing using an identical, optimized training protocol on two training systems under three listening conditions (open ear, TEP-100, and ComTac™ III). Statistical tests were performed on mean absolute accuracy score and front-back reversal errors. No statistical difference existed between the PALAT and laboratory-grade DRILCOM systems on two dependent localization accuracy measurements at all stages of training. In addition, the PALAT system detected the same localization performance differences among the three listening conditions. The PALAT system imparted similar training benefits as the DRILCOM system and was sensitive to HPD localization performance differences. The user-operable PALAT system and optimized training protocol can be employed by the military, law enforcement, and various industries, to improve auditory localization performance in conditions where auditory situation awareness is critical to safety.

Multi-Output Regression Indoor Localization Algorithm Based on Hybrid Grey Wolf Particle Swarm Optimization

In the evolving landscape of device-free localization techniques, Wi-Fi channel state information (CSI) emerges as a pivotal tool for environmental sensing. This study introduces a novel fingerprint localization algorithm. It employs an improved Hybrid Grey Wolf Particle Swarm Optimization (IPSOGWO) in combination with Multi-Output Support Vector Regression (MSVR) to enhance indoor positioning accuracy. To counteract the limitations of standard DBSCAN and PCA in noise reduction and feature extraction from complex nonlinear data, we propose an adaptive denoising algorithm based on spatial clustering (A-DBSCAN) and an autoencoder to efficiently denoise and extract features from CSI amplitude to improve the localization accuracy. Additionally, we introduce a new position update strategy, bolstering the optimization efficiency of the PSOGWO algorithm. This refined approach is instrumental in determining the globally optimal hyperparameters in MSVR, leading to enhanced model prediction accuracy. Two indoor scenario experiments were conducted to evaluate our method, yielding average localization errors of 0.59 m and 1.12 m, marking an improvement in localization performance compared to existing methods.

Case Sampling vs Universal Review for Evaluating Hospital Postoperative Mortality in US Surgical Quality Improvement Programs

Representative surgical case sampling, rather than universal review, is used by US Department of Veterans Affairs (VA) and private-sector national surgical quality improvement (QI) programs to assess program performance and to inform local QI and performance improvement efforts. However, it is unclear whether case sampling is robust for identifying hospitals with safety or quality concerns. To evaluate whether the sampling strategy used by several national surgical QI programs provides hospitals with data that are representative of their overall quality and safety, as measured by 30-day mortality. This comparative effectiveness study was a national, hospital-level analysis of data from adult patients (aged ≥18 years) who underwent noncardiac surgery at a VA hospital between January 1, 2016, and September 30, 2020. Data were obtained from the VA Surgical Quality Improvement Program (representative sample) and the VA Corporate Data Warehouse surgical domain (100% of surgical cases). Data analysis was performed from July 1 to December 21, 2022. The primary outcome was postoperative 30-day mortality. Quarterly, risk-adjusted, 30-day mortality observed-to-expected (O-E) ratios were calculated separately for each hospital using the sample and universal review cohorts. Outlier hospitals (ie, those with higher-than-expected mortality) were identified using an O-E ratio significantly greater than 1.0. In this study of data from 113 US Department of Veterans Affairs hospitals, the sample cohort comprised 502 953 surgical cases and the universal review cohort comprised 1 703 140. The majority of patients in both the representative sample and the universal sample were men (90.2% vs 91.1%) and were White (74.7% vs 74.5%). Overall, 30-day mortality was 0.8% and 0.6% for the sample and universal review cohorts, respectively (P < .001). Over 2145 quarters of data, hospitals were identified as an outlier in 11.7% of quarters with sampling and in 13.2% with universal review. Average hospital quarterly 30-day mortality rates were 0.4%, 0.8%, and 0.9% for outlier hospitals identified using the sample only, universal review only, and concurrent identification in both data sources, respectively. For nonsampled cases, average hospital quarterly 30-day mortality rates were 1.0% at outlier hospitals and 0.5% at nonoutliers. Among outlier hospital quarters in the sample, 47.4% were concurrently identified with universal review. For those identified with universal review, 42.1% were concurrently identified using the sample. In this national, hospital-level study, sampling strategies employed by national surgical QI programs identified less than half of hospitals with higher-than-expected perioperative mortality. These findings suggest that sampling may not adequately represent overall surgical program performance or provide stakeholders with the data necessary to inform QI efforts.

Performance characteristics on up-scaling vacuum membrane distillation modules

Large-scale CFD studies are currently lacking in membrane distillation due to high computational overhead. This study utilizes a validated CFD model for vacuum membrane distillation and reduces the order of the model to enable prototype scale simulations. Different module designs were considered: a baseline flat sheet membrane module, a spacer-filled channel, a wavy channel, and a wavy channel with embedded stiffeners. Different lengths (330 mm, 660 mm, and 990 mm) were studied to simulate lab to prototype upscaling. Large Eddy Simulations were run in the laminar (Re 900) and turbulent (Re 3600) flow regimes. For Re 900, cases with geometric alterations (wavy membrane, spacer-filled channel) locally outperformed the baseline case within the first 330 mm before axial temperature polarization equilibrated the performance between the baseline and altered channels. The improvement in local performance was seen over the whole 990 mm length at Re 3600, culminating in a 35–40 % flux improvement versus the baseline at Re 3600 versus a 0–35 % difference at Re 900. Axial temperature polarization was mitigated using heat injection, showing a 5 % flux drop as length tripled compared to a 20–27 % decline. Higher operational Reynolds numbers should be considered cautiously for increasing pressure drop and violating the liquid entry pressure limit for wetting.

TransBoNet: Learning camera localization with Transformer Bottleneck and Attention

6DoF camera localization is an important component of autonomous driving and navigation. Deep learning has achieved impressive results in localization, but its robustness in dynamic environments has not been adequately addressed. In this paper, we propose a framework based on hybrid attention mechanism which can be generally applied to existing CNN-based pose regressors to improve their robustness in dynamic environments. Specifically, we propose a novel Transformer Bottleneck (TBo) block including convolution, channel attention, and a position-aware self-attention mechanism, which extracts more geometrically robust features by capturing the corresponding long-term dependencies between pixels. Furthermore, we introduce shuffle attention (SA) before the pose regressor, which integrates feature information in both spatial and channel dimensions, forcing the network to learn geometrically robust features, reducing the effects of dynamic objects and illumination conditions to improve camera localization accuracy. We evaluate our method on commonly benchmarked indoor and outdoor datasets and the experimental results show that our proposed method can significantly improve localization performance compared compare favorably to contemporary pose regressors schemes. In addition, extensive ablation evaluations are conducted to prove the effectiveness of our proposed hybrid attention bottleneck block for pose regression networks.

A method for localization using network and analysis of data error

Facing the problem of radiator target localization in a specific airspace, this paper puts forward the idea of time difference of arrival (TDOA) localization using the neural network. In the paper, we build a TDOA model for targets in an airspace to obtain the position and the TDOA value pairs, and then use them for a large amount of network training. The trained network can perform real-time localization. Furthermore, the paper also deduces the rationality of the error contained in the TDOA data in the training model, which proves the improvement of network performance in this case. Finally, the coefficient of determination is introduced to measure the training effect of the network. The experimental results verify the feasibility of using neural network for localization, and it can obtain better effectiveness and robustness compared with traditional algorithms. Also, a large number of error experiments are used to verify that training data needs to be added with appropriate errors based on actual localization scenarios to improve localization performance.

Adaptive Scheme of Denoising Autoencoder for Estimating Indoor Localization Based on RSSI Analytics in BLE Environment

In indoor environments, estimating localization using a received signal strength indicator (RSSI) is difficult because of the noise from signals reflected and refracted by walls and obstacles. In this study, we used a denoising autoencoder (DAE) to remove noise in the RSSI of Bluetooth Low Energy (BLE) signals to improve localization performance. In addition, it is known that the signal of an RSSI can be exponentially aggravated when the noise is increased proportionally to the square of the distance increment. Based on the problem, to effectively remove the noise by adapting this characteristic, we proposed adaptive noise generation schemes to train the DAE model to reflect the characteristics in which the signal-to-noise ratio (SNR) considerably increases as the distance between the terminal and beacon increases. We compared the model’s performance with that of Gaussian noise and other localization algorithms. The results showed an accuracy of 72.6%, a 10.2% improvement over the model with Gaussian noise. Furthermore, our model outperformed the Kalman filter in terms of denoising.

Versatile LiDAR-Inertial Odometry With SE(2) Constraints for Ground Vehicles

LiDAR SLAM has become one of the major localization systems for ground vehicles since LiDAR Odometry And Mapping (LOAM). Many extension works on LOAM mainly leverage one specific constraint to improve the performance, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e.g.</i> , information from on-board sensors such as loop closure and inertial state; prior conditions such as ground level and motion dynamics. In many robotic applications, these conditions are often known partially, hence a SLAM system can be a comprehensive problem due to the existence of numerous constraints. Therefore, we can achieve a better SLAM result by fusing them properly. In this paper, we propose a hybrid LiDAR-inertial SLAM framework that leverages both the on-board perception system and prior information such as motion dynamics to improve localization performance. In particular, we consider the case for ground vehicles, which are commonly used for autonomous driving and warehouse logistics. We present a computationally efficient LiDAR-inertial odometry method that directly parameterizes ground vehicle poses on SE(2). The out-of-SE(2) motion perturbations are not neglected but incorporated into an integrated noise term of a novel SE(2)-constraints model. For odometric measurement processing, we propose a versatile, tightly coupled LiDAR-inertial odometry to achieve better pose estimation than traditional LiDAR odometry. Thorough experiments are performed to evaluate our proposed method's performance in different scenarios, including localization for both indoor and outdoor environments. The proposed method achieves superior performance in accuracy and robustness.

Improved Discriminative Object Localization Algorithm for Safety Management of Indoor Construction.

Object localization is a sub-field of computer vision-based object recognition technology that identifies object classes and locations. Studies on safety management are still in their infancy, particularly those aimed at lowering occupational fatalities and accidents at indoor construction sites. In comparison to manual procedures, this study suggests an improved discriminative object localization (IDOL) algorithm to aid safety managers with visualization to improve indoor construction site safety management. The IDOL algorithm employs Grad-CAM visualization images from the EfficientNet-B7 classification network to automatically identify internal characteristics pertinent to the set of classes evaluated by the network model without the need for further annotation. To evaluate the performance of the presented algorithm in the study, localization accuracy in 2D coordinates and localization error in 3D coordinates of the IDOL algorithm and YOLOv5 object detection model, a leading object detection method in the current research area, are compared. The comparison findings demonstrate that the IDOL algorithm provides a higher localization accuracy with more precise coordinates than the YOLOv5 model over both 2D images and 3D point cloud coordinates. The results of the study indicate that the IDOL algorithm achieved improved localization performance over the existing YOLOv5 object detection model and, thus, is able to assist with visualization of indoor construction sites in order to enhance safety management.

IWFUCIA: An integrated model for Wi-Fi fingerprints radio map updating and localization on changed impact analysis

In Wi-Fi fingerprinting indoor localization, automating radio map database maintenance is one of the crucial issues, as it is a labour-intensive and long-term task for collecting and filtering samples to keep an up-to-date and accurate database. In particular, those access points (APs) newly installed in the environment should update radio maps and be included in the database to improve localization performance. This study presents an IWFUCIA system that automates indoor radio map database maintenance (RMapDM) using crowdsourced samples without accurate location annotation. The IWFUCIA incorporates the newly installed APs detection and identification, the significant APs feature selection, fingerprint integration updating, and online localization algorithms. After collecting new crowdsourced samples, we apply Willmott’s index of agreement (WIA) based on the Supported Vector Machine (SVM) regression to detect and identify a newly installed AP and the original existing ones. After getting the new APs, we propose a correlated coefficient and t-test score algorithm to select only those significant AP-based feature samples. We also proposed a fingerprint integration model to fuse original existing and new APs to update the database. Extensive experiments have been conducted in our teaching building to validate and evaluate the effectiveness of IWFUCIA. The results show that our IWFUCIA is robust for long-term maintenance and updating the outdated radio map database server. The average localization accuracy achieves 0.466 m, which significantly outperforms the localization positioning approaches with the original radio map by 84.96%, outdated radio maps by the changed APs powers removed, increased and decreased by 26.32%, 55.36%, and 73.14%, respectively.