Target Tracking Research Articles

Rapid Response of Daphnia magna Motor Behavior to Mercury Chloride Toxicity Based on Target Tracking.

A rapid and timely response to the impacts of mercury chloride, which is indispensable to the chemical industry, on aquatic organisms is of great significance. Here, we investigated whether the YOLOX (improvements to the YOLO series, forming a new high-performance detector) observation system can be used for the rapid detection of the response of Daphnia magna targets to mercury chloride stress. Thus, we used this system for the real-time tracking and observation of the multidimensional motional behavior of D. magna. The results obtained showed that the average velocity (v¯), average acceleration (a¯), and cumulative travel (L) values of D. magna exposed to mercury chloride stress changed significantly under different exposure times and concentrations. Further, we observed that v¯, a¯ and L values of D. magna could be used as indexes of toxicity response. Analysis also showed evident D. magna inhibition at exposure concentrations of 0.08 and 0.02 mg/L after exposure for 10 and 25 min, respectively. However, under 0.06 and 0.04 mg/L toxic stress, v¯ and L showed faster toxic response than a¯, and overall, v¯ was identified as the most sensitive index for the rapid detection of D. magna response to toxicity stress. Therefore, we provide a strategy for tracking the motile behavior of D. magna in response to toxic stress and lay the foundations for the comprehensive screening of toxicity in water based on motile behavior.

A transfer learning model for cognitive electronic reconnaissance of unmanned aerial vehicle: Experiments

Applying Deep Reinforcement Learning (DRL) technologies to Unmanned Aerial Vehicle (UAV) electronic reconnaissance is one of the current research hotspots. However, simulation and engineering practice show that due to poor generalization of DRL models, the performance of Cognitive Electronic Reconnaissance (CER) policies based on training will significantly decrease when the mission scene undergoes slight changes. To address this issue, we thoughtfully combine the mission area segmentation technique with transfer DRL and propose a difference-adaptive transfer DRL algorithm. This algorithm involves mission subarea segmentation, subarea pre-training, multi-subarea policy transfer, and multi-subarea splicing, providing an efficient solution to the convergence problem of the DRL algorithm caused by mission space expansion and reward sparsity. Additionally, a general CER transfer learning simulator is developed based on the analysis of the capabilities of the maneuvering platform and electronic reconnaissance payload. Multiple sets of CER policy transfer learning experiments are designed for different mission spaces, mission difficulties, and UAV characteristics. Compared with the algorithm baseline, our designed policy model significantly outperforms: the mission completion rate of UAVs in multi-scale mission spaces improves by up to 37.4%, reaching 97.5%, while the training time is reduced by 2.46 h. Further behavior analysis shows that this policy model enables UAVs to exhibit target tracking behaviors such as hovering and sustained approach.

Siamese Tracking Network with Multi-attention Mechanism

Object trackers based on Siamese networks view tracking as a similarity-matching process. However, the correlation operation operates as a local linear matching process, limiting the tracker’s ability to capture the intricate nonlinear relationship between the template and search region branches. Moreover, most trackers don’t update the template and often use the first frame of an image as the initial template, which will easily lead to poor tracking performance of the algorithm when facing instances of deformation, scale variation, and occlusion of the tracking target. To this end, we propose a Simases tracking network with a multi-attention mechanism, including a template branch and a search branch. To adapt to changes in target appearance, we integrate dynamic templates and multi-attention mechanisms in the template branch to obtain more effective feature representation by fusing the features of initial templates and dynamic templates. To enhance the robustness of the tracking model, we utilize a multi-attention mechanism in the search branch that shares weights with the template branch to obtain multi-scale feature representation by fusing search region features at different scales. In addition, we design a lightweight and simple feature fusion mechanism, in which the Transformer encoder structure is utilized to fuse the information of the template area and search area, and the dynamic template is updated online based on confidence. Experimental results on publicly tracking datasets show that the proposed method achieves competitive results compared to several state-of-the-art trackers.

Adaptive horizon size moving horizon estimation with unknown noise statistical properties

Abstract Moving horizon estimation (MHE) is an effective technique for state estimation. It formulates state estimation as an optimization problem over a finite time interval and is characterized by inherent robustness, flexibility, and explicit constraint handling capabilities. The horizon size is a crucial parameter influencing the estimation performance of MHE. However, the selection of the horizon size remains an open research question in the field of MHE. In this paper, we propose a novel adaptive horizon size MHE strategy that dynamically adjusts the horizon size based on the value of the objective function. This approach aims to improve the state estimation performance of MHE in real-time applications. Unlike conventional MHE methods that rely on a fixed horizon size, our adaptive strategy enhances robustness against unknown noise statistics by adjusting the horizon size. We analyze the convergence property of the estimation error and provide guidelines for parameter design to ensure optimal performance. The effectiveness and superiority of the proposed method are demonstrated through simulations involving an oscillatory system and a target tracking application under non-stationary noise conditions.

Hyperspectral image restoration using noise gradient and dual priors under mixed noise conditions

AbstractImages obtained from hyperspectral sensors provide information about the target area that extends beyond the visible portions of the electromagnetic spectrum. However, due to sensor limitations and imperfections during the image acquisition and transmission phases, noise is introduced into the acquired image, which can have a negative impact on downstream analyses such as classification, target tracking, and spectral unmixing. Noise in hyperspectral images (HSI) is modelled as a combination from several sources, including Gaussian/impulse noise, stripes, and deadlines. An HSI restoration method for such a mixed noise model is proposed. First, a joint optimisation framework is proposed for recovering hyperspectral data corrupted by mixed Gaussian‐impulse noise by estimating both the clean data as well as the sparse/impulse noise levels. Second, a hyper‐Laplacian prior is used along both the spatial and spectral dimensions to express sparsity in clean image gradients. Third, to model the sparse nature of impulse noise, an ℓ1 − norm over the impulse noise gradient is used. Because the proposed methodology employs two distinct priors, the authors refer to it as the hyperspectral dual prior (HySpDualP) denoiser. To the best of authors' knowledge, this joint optimisation framework is the first attempt in this direction. To handle the non‐smooth and non‐convex nature of the general ℓp − norm‐based regularisation term, a generalised shrinkage/thresholding (GST) solver is employed. Finally, an efficient split‐Bregman approach is used to solve the resulting optimisation problem. Experimental results on synthetic data and real HSI datacube obtained from hyperspectral sensors demonstrate that the authors’ proposed model outperforms state‐of‐the‐art methods, both visually and in terms of various image quality assessment metrics.

MIMTracking: Masked image modeling enhanced vision transformer for visual object tracking

Recently, one-stage trackers achieve state-of-the-art tracking results due to more sufficient integration of search and template representations. These methods usually adopt an encoder for synchronous feature generation and interaction. Despite their high performance, the one-stage approaches tend to feed the encoder full input representations that are highly redundant and dense during training. With a focus on tackling this problem, a novel algorithm termed MIMTracking is developed for visual target tracking. MIMTracking exploits an encoder and a decoder for masked image modeling during training. Randomly sampled discrete search embeddings and template embeddings serve as input of the encoder. The lightweight decoder takes full representations as input and progressively highlights the target region. This design alleviates input redundancy and reduces the computational cost of the training process, thereby allowing for more efficient learning of useful representations. The proposed MIMTracking achieves state-of-the-art tracking results on numerous tracking datasets, e.g., 51.5 % area under curve (AUC) on LaSOText, outperforming the previous top tracker OSTrack by 2 %. Especially, our large model MIMTracking-L further improves the AUC to 53.4 % on LaSOText.

Design and application of UAV single target tracking system based on YOLOv5

In aerial reconnaissance, pursuit, and post-disaster search and rescue, rotorcraft UAVs are valued for their agility. A key challenge is target tracking in complex environments with no prior information, dense obstacles, GNSS signal loss, and occluded targets. This paper introduces a YOLOv5-based UAV tracking system to address detection and path planning. The quadrotor UAV model ensures stable flight through mathematical modeling and simulation. YOLOv5 enhances detection performance with real-time tracking. Feedback control enables autonomous obstacle avoidance and target tracking. Results show adaptability and robustness, supporting reconnaissance and rescue missions effectively.

Performance modeling of composite fading channels: an application to wearable wireless communication systems

In pragmatic wireless propagation scenarios, communication between wearable devices may be significantly impacted by fading, particularly when wireless communication is crucial for mission control, situational awareness, and team coordination in defense operations. There may be times when communication links are temporarily unavailable in places due to contemporaneous effect of slow fading and fast fading, including cities or regions with tough terrain. This phenomena results in composite fading environments. Moreover, the data transfer speeds may drop as a result of fading, making it a challenging task to rapidly transmit crucial real time data. In addition, the precision of GPS and positioning systems, that are essential for operations viz., target tracking, navigation and rescue can be impacted by fading. In this setting, the work portrays modeling the fading channels over the Weibull distribution by exploiting Tsallis’ q-statistical framework. From the analytical observations, it was evident that in contrast to the well-known composite models, the q-Weibull model provides a better characterization of the generated fading signals. Furthermore, the closed form solution to the different performance metrics that are crucial in characterizing the fading environments are provided.

Bearing-driven approach for unknown target tracking by multi-unit uncertain underactuated marine autonomous surface vehicles

In this study, we successfully address the formation tracking challenge for underactuated autonomous surface vehicles (ASVs) in uncertain environments, despite the lack of direct target information. Leveraging only relative angle measurements, we develop a control scheme that ensures the closed-loop error trajectory remains globally uniformly ultimately bounded. The bound achieved is influenced by the precision of the dynamic controllers and the effectiveness of the adaptive mechanism, highlighting the robustness and adaptability of our approach. This work provides a practical solution for ASV formation tracking in real-world scenarios.

Precision of trunk movement in people with chronic low back pain

BackgroundMotor control exercise is commonly applied in people with chronic low back pain (CLBP), but possibly not all people with CLBP have motor control impairments. We suggest movement precision as measure to identify motor control impairments. Movement precision has been operationalized as trunk movement variability (TMV) and as trunk tracking error(s) (TTE). Objectives: To compare the known-group validity and the responsiveness of TMV and TTE. DesignWe used a case-control comparison (Healthy controls (n = 30) vs CLBP (n = 60)) to assess the known-group validity. A cohort study, (measurements in week 3 and week 12 of treatment), was used to assess responsiveness. MethodsTMV (temporal (CyclSD) and spatial (MeanSD)) was analyzed during standing, repetitive flexion and rotation tasks (35x). TTE was measured during movement target tracking tasks, again in flexion and rotation. Participants with CLBP followed a multidisciplinary intervention and both measures were assessed in week 3 and week 12 of treatment. 2-way MANOVA and 2-way ANOVA were used to assess the effect of Group (CLBP vs healthy controls) and direction (flexion vs rotation) on TMV and TTE. For responsiveness, 2-way MANOVA and 2-way ANOVA were used to assess the effect of treatment and direction on both measures. FindingsAt baseline, TMV was not different between groups, while TTE was higher in the people with CLBP (p = 0.005, np2 = 0.09). Treatment strongly decreased temporal TMV (p = 0.025, np2 = 0.33) and TTE (p < 0.001, np2 = 0.844). ConclusionsThese results demonstrate that TTE is more sensitive to CLBP and more responsive to treatment than TMV.

Maneuvering extended target tracking method based on transformer network

This research introduces a new approach to address the challenges related to inaccuracies in shape estimation and motion state tracking during the tracking of maneuvering extended targets. The method combines a novel neural network model with a cubature Kalman filter. Initially, the target’s shape variation is represented using a second-order autoregressive model, and Bayesian filtering is employed for target contour estimation in static environments. Subsequently, the kinematic state tracking of maneuvering targets is enhanced by modifying a neural network model based on Transformer architecture. Through the effective integration of these techniques, precise tracking of maneuvering extended targets is achieved, facilitating accurate estimation of irregular contour information while tracking the targets’ motion state in real-time. The efficacy of the proposed approach is further confirmed through various simulation scenarios, underscoring its suitability for tracking maneuvering extended targets.

Application of fuzzy logic control theory combined with target tracking algorithm in unmanned aerial vehicle target tracking

This paper aims to increase the Unmanned Aerial Vehicle's (UAV) capacity for target tracking. First, a control model based on fuzzy logic is created, which modifies the UAV's flight attitude in response to the target's motion status and changes in the surrounding environment. Then, an edge computing-based target tracking framework is created. By deploying edge devices around the UAV, the calculation of target recognition and position prediction is transferred from the central processing unit to the edge nodes. Finally, the latest Vision Transformer model is adopted for target recognition, the image is divided into uniform blocks, and then the attention mechanism is used to capture the relationship between different blocks to realize real-time image analysis. To anticipate the position, the particle filter algorithm is used with historical data and sensor inputs to produce a high-precision estimate of the target position. The experimental results in different scenes show that the average target capture time of the algorithm based on fuzzy logic control is shortened by 20% compared with the traditional proportional-integral-derivative (PID) method, from 5.2 s of the traditional PID to 4.2 s. The average tracking error is reduced by 15%, from 0.8 m of traditional PID to 0.68 m. Meanwhile, in the case of environmental change and target motion change, this algorithm shows better robustness, and the fluctuation range of tracking error is only half of that of traditional PID. This shows that the fuzzy logic control theory is successfully applied to the UAV target tracking field, which proves the effectiveness of this method in improving the target tracking performance.

Template‐Refine Network for Siamese Object Tracking

Various mainstream target tracking algorithms based on Siamese networks are gradually becoming a trend in the field of deep learning tracking due to their concurrent advantages of accuracy and speed. Most Siamese network‐based trackers describe the tracking of a target object as a similar matching problem, and these trackers have achieved more advanced performance in several public tests. Most trackers often suffer from tracking drift or performance degradation owing to the non‐updating of the template in the first frame and the target appearance encounters disturbing environments such as occlusion and drastic deformation. Therefore, to address this problem, this paper introduces a template updating mechanism and proposes a refine structure network based on the template updating of Siamese networks as well as the greater similarity of target features in two adjacent frames, which improves the tracking accuracy while limiting the amount of computation using an anchor‐free method in order not to lose the tracking speed, and only needs to be trained by selecting the most suitable pre‐training network, thus greatly reducing the amount of network computation. Meanwhile, in the application of the refine structure, with the aim of making the weight design of the target localisation module more reasonable, we propose a new Refine Head section and analyze and design the update threshold to optimize the overall network. This method is practiced in SiamFC++ algorithm, which firstly designs the template refine module, inputs the image that needs to be improved, and then outputs it to the Refine Head to complete the template update and applies it to the tracking of the subsequent frames, thereby constituting the SiamTRN (Template‐Refine Network). According to the experiments, the improved structure of the method can effectively implement the refine module function and enhance the performance of the tracker on public datasets, such as OTB100, VOT2016, UAV123 and GOT‐10 k. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Natural Language Processing Pretraining Language Model for Computer Intelligent Recognition Technology

Computer intelligent recognition technology refers to the use of computer vision, Natural Language Processing (NLP), machine learning and other technologies to enable computers to recognize, analyze, understand and answer human language and behavior. The common applications of computer intelligent recognition technology include image recognition, NLP, face recognition, target tracking, and other fields. NLP is a field of computer science, which involves the interaction between computers and natural languages. NLP technology can be used to process, analyze and generate natural language data, such as text, voice and image. Common NLP technology applications include language translation, emotion analysis, text classification, speech recognition and question answering system. Language model is a machine learning model, which uses a large number of text data for training to learn language patterns and relationships in text data. Although the language model has made great progress in the past few years, it still faces some challenges, including: poor semantic understanding, confusion in multilingual processing, slow language processing and other shortcomings. Therefore, in order to optimize these shortcomings, this article would study the pre-training language model based on NLP technology, which aimed at using NLP technology to optimize and improve the performance of the language model, thus optimizing the computer intelligent recognition technology. The model had a higher language understanding ability and more accurate prediction ability. In addition, the model could learn language rules and structures by using a large number of corpus, so as to better understand natural language. Through experiments, it could be known that the data size and total computing time of the traditional Generative Pretrained Transformer-2 (GPT-2) language model were 10 GB and 97 hours respectively. The data size and total computing time of BERT (Bidirectional Encoder Representations from Transformer) were 12 GB and 86 hours respectively. The data size and total computing time of the pre-training language model based on NLP were 18 GB and 71 hours respectively. Obviously, the pre-training language model based on NLP had a larger data size and shorter computing time. The experimental results showed that the NLP technology could better optimize the language model and effectively improve its various capabilities. This article opened up a new development direction for computer intelligent recognition technology and provided excellent technical support for the development of language models.

Optical Detection of Underwater Propeller Wake Based on a Position-Sensitive Detector

The study of underwater vehicle wake detection is of significant importance within the field of target detection, localisation, and tracking of underwater vehicles. Given that propellers are the propellers of modern ships and underwater vehicles, the propeller wake field represents the principal target source for wake detection in underwater vehicles. The objective of this paper is to propose a method for measuring the wake of an underwater propeller based on a position-sensitive detector. A theoretical model of the relationship between the laser spot displacement and the change in the refractive index of the wake field is established on the basis of the principle of laser beam deflection. A prototype experimental setup for underwater propeller wake measurement was constructed based on the aforementioned optical measurement method. Furthermore, the simulation of the propeller wake flow field with strong density stratification and linear density stratification was conducted based on the experimental setup. Furthermore, experiments were conducted to detect the flow field of a propeller wake. The experimental results indicate that the wake dissipation times of the propeller in a strong density-stratified water environment are approximately 800 s and 750 s. Following the stabilisation of the wake field density, the laser spot position is observed to be stable at 0.341 mm and 0.441 mm, respectively, with a corresponding refractive index change of 2.99 × 10−6 RIU (refractive index unit) and 3.87 × 10−6 RIU, respectively. These experimental results are found to be in general agreement with the simulation results of the propeller wake field. A comparison of the experimental wake measurements based on the device with the wake measurements based on a CTD (conductivity–temperature–depth) device reveals a consistent trend. The realisation of this detection technique is of great significance for the advancement of research in the field of optical detection of underwater vehicle wake streams.

Performance Improvement Approach to Naval Gun Fire Control System Based on Linear Target Tracking Filter with Radar Line-of-sight Measurements

This paper addresses a novel approach to performance enhancement of the naval gun fire control system(FCS) by using the projectile tracking filter without any distortion of radar measurements. Under the assumption that the maneuvering between the projectile and the ship equipped with the radar is not quite large, this method is based on the concept of polar-coordinate target tracking, which separates the range estimation filter and the direction cosine estimation filter. Note that using polar-coordinates allows tracking to be performed in the same coordinate system from which the radar line-of-sight(LOS) measurements are obtained, unlike the conventional tracking process in Cartesian. Also, it is easy to implement in real-time and guarantees consistent estimates due to its linear filter structure. With the help of the above method, therefore, the proposed filter is able to improve the overall performance of FCS which requires stability of projectile estimates within a short engagement time. The effectiveness of the presented scheme is validated through computer simulations.

A Real-Time Power Sharing Strategy for a Multi-Energy System While Considering Individual System Dynamics

For a sustainable and resilient energy supply, multi-energy systems (MES) are becoming more prominent. Real-time efficient power-sharing in an MES consisting of different generation sources, uncontrollable loads, multiple storage options, and Power-to-X technologies is a challenging task. The challenge arises due to constantly fluctuating generation and load, as well as the different system dynamics of each element in the setup. Optimizing the power-sharing among the various controllable elements in an MES could be broken down into a two-level process. The top-level process, commonly known as super-ordinate control, defines the power-sharing over a longer timeframe based on load and generation forecasts and multiple other factors. The lower-level process, commonly known as sub-ordinate control, based on the inputs from the super-ordinate control and live sensor data refines the control signals of the individual elements. This work focuses on developing a control framework for the sub-ordinate control process while considering the individual element dynamics such as dead time and rise time to a control signal. This is important as not considering differences in the system dynamics results in sub-optimal control, causing dynamic mismatches. This work proposes a simple rule-based power-splitting method backed up with individual PID regulators and Smith Predictor for each element in the MES, which are further coupled to each other for more precise and efficient control. The proposed method induces a cooperative behavior among the MES elements and improves the reaction time to a control signal while also improving the target tracking by 3-8%.

Application of Video Abnormal Behavior Detection Algorithm in Evaluation of Track and Field Teaching and Training Effect

With the development of track and field, people pay more and more attention to the quality of classroom teaching of track and field technology, and the evaluation of teaching quality plays a key role in it. In today's educational reform, teaching evaluation plays an important role as an important method to test teachers' teaching and students' learning. With the rapid development of machine learning, especially deep learning, image-based individual abnormal behavior detection technology is becoming more and more mature, but there are still many difficulties to be solved in video-based group abnormal behavior detection technology. Therefore, it is necessary to study the detection algorithm. Based on machine vision theory, image processing theory and video analysis technology, this paper studies three key technologies involved in human abnormal behavior detection in video: moving target detection, moving target tracking and abnormal behavior detection.

Application of image recognition-based tracker-less augmented reality navigation system in a series of sawbone trials

BackgroundThis study introduced an Augmented Reality (AR) navigation system to address limitations in conventional high tibial osteotomy (HTO). The objective was to enhance precision and efficiency in HTO procedures, overcoming challenges such as inconsistent postoperative alignment and potential neurovascular damage.MethodsThe AR-MR (Mixed Reality) navigation system, comprising HoloLens, Unity Engine, and Vuforia software, was employed for pre-clinical trials using tibial sawbone models. CT images generated 3D anatomical models, projected via HoloLens, allowing surgeons to interact through intuitive hand gestures. The critical procedure of target tracking, essential for aligning virtual and real objects, was facilitated by Vuforia’s feature detection algorithm.ResultsIn trials, the AR-MR system demonstrated significant reductions in both preoperative planning and intraoperative times compared to conventional navigation and metal 3D-printed surgical guides. The AR system, while exhibiting lower accuracy, exhibited efficiency, making it a promising option for HTO procedures. The preoperative planning time for the AR system was notably shorter (4 min) compared to conventional navigation (30.5 min) and metal guides (75.5 min). Intraoperative time for AR lasted 8.5 min, considerably faster than that of conventional navigation (31.5 min) and metal guides (10.5 min).ConclusionsThe AR navigation system presents a transformative approach to HTO, offering a trade-off between accuracy and efficiency. Ongoing improvements, such as the incorporation of two-stage registration and pointing devices, could further enhance precision. While the system may be less accurate, its efficiency renders it a potential breakthrough in orthopedic surgery, particularly for reducing unnecessary harm and streamlining surgical procedures.

Graph-based multi-agent reinforcement learning for collaborative search and tracking of multiple UAVs

This paper investigates the challenges associated with Unmanned Aerial Vehicle (UAV) collaborative search and target tracking in dynamic and unknown environments characterized by limited field of view. The primary objective is to explore the unknown environments to locate and track targets effectively. To address this problem, we propose a novel Multi-Agent Reinforcement Learning (MARL) method based on Graph Neural Network (GNN). Firstly, a method is introduced for encoding continuous-space multi-UAV problem data into spatial graphs which establish essential relationships among agents, obstacles, and targets. Secondly, a Graph AttenTion network (GAT) model is presented, which focuses exclusively on adjacent nodes, learns attention weights adaptively and allows agents to better process information in dynamic environments. Reward functions are specifically designed to tackle exploration challenges in environments with sparse rewards. By introducing a framework that integrates centralized training and distributed execution, the advancement of models is facilitated. Simulation results show that the proposed method outperforms the existing MARL method in search rate and tracking performance with less collisions. The experiments show that the proposed method can be extended to applications with a larger number of agents, which provides a potential solution to the challenging problem of multi-UAV autonomous tracking in dynamic unknown environments.