Improve Tracking Performance Research Articles

Observability-Based Gaussian Sum Cubature Kalman Filter for Three-Dimensional Target Tracking Using a Single Two-Dimensional Radar

This paper considers the problem of tracking a three-dimensional target under the condition that only a single two-dimensional radar is available. Since a two-dimensional radar can only measure the slant range and azimuth information relative to the target, an unobservability issue arises in this tracking application. Therefore, we first investigate the observability issue of tracking a three-dimensional target with a single two-dimensional radar from two perspectives, including intuitive illustration and quantitative analysis. From the perspective of intuitive illustration, we demonstrate “What is the unobservability issue” and “How does the relative target-radar geometry influence the observability of the tracking system”. From the perspective of quantitative analysis, we construct a novel observability metric for this special tracking problem. Second, aiming at improving tracking performance under the unobservability of target height, we propose an observability-based Gaussian sum cubature Kalman filter. Built within the Gaussian sum framework and based on the height-parameterized strategy, this novel algorithm uses a set of independent fifth-degree cubature Kalman filters, each of which can detect the system observability variation and enhance the tracking accuracy by using a Gaussian splitting scheme under low-degree observability. Finally, the effectiveness of the presented filtering algorithm is validated through lots of simulation experiments.

Improvement of tracking performance of SMA-actuated rotary actuator via thermocouple delay compensation using adaptive predictor

This paper investigates the impact of bounded and unknown delay of a J-type thermocouple sensor on controlling a Shape Memory Alloy (SMA) actuated rotary actuator. To compensate for the delay, a nonlinear adaptive predictor is utilized, which predicts the exact temperature of the present moment using the output of the sensor. A robust sliding mode control is then designed to track the reference input, which is the rotation angle of the pulley. In addition to temperature estimation, the inner system identification algorithm of the predictor identifies the heat transfer coefficients of the experimental setup. Experimental tests demonstrate the precise control of the proposed algorithm. The tracking error of the adaptive predictor with sliding mode control is compared to that of a regular sliding mode controller without temperature delay compensation consideration, showing the superior performance of the proposed control algorithm.

An Adaptive Sliding Mode Control Using a Novel Adaptive Law Based on Quasi-Convex Functions and Average Sliding Variables for Robot Manipulators

This paper proposes a novel adaptive law that uses a quasi-convex function and a novel sliding variable in an adaptive sliding mode control (ASMC) scheme for robot manipulators. Since the dynamic equations of robot manipulators inevitably include model uncertainties and disturbances, time-delay estimation (TDE) errors occur when using the time-delay control (TDC) approach. Further, the ASMC method used to compensate for TDE errors naturally causes a chattering phenomenon. To improve tracking performance while reducing or maintaining chattering, this paper proposes an adaptive law based on a quasi-convex function that is convex at the origin and concave at the gain switching point, respectively. We also adopt a novel sliding variable that uses previously sampled tracking errors and their time derivatives. Further, this paper proves that the sliding variable of the robot manipulator controlled by the proposed ASMC satisfies uniformly ultimately bounded stability. The simulation and experimental results illustrate the effectiveness of the proposed methods in terms of tracking performance.

Sensorless Human–Robot Interaction: Real‐Time Estimation of Co‐Grasped Object Mass and Human Wrench for Compliant Interaction

Human–robot physical interaction in shared object manipulation can fully leverage the strengths of both humans and collaborative robots, achieving better interaction outcomes. However, in typical physical interactions, the inertia parameters of the object are either assumed to be known or ignored, simplifying the interaction to be directly between the human and the collaborative robot. This simplification can lead to inaccuracies in the robot's understanding of human intent due to deviations in object dynamics. This article presents a method that eliminates the need for a force sensor on the human side during human–robot collaboration. Using an extended Kalman filter, the method estimates object parameters online and applies a disturbance observer to determine human force. This approach reduces human effort and improves tracking performance compared to traditional methods. It also avoids the hassle of installing and removing sensors in complex scenarios. By estimating object mass and human force in real time, the robot can adjust its behavior to ensure safer and smoother interactions. Furthermore, the system can handle various objects without requiring specific programming for each case.

Long-term tracking for high-frequency surface wave radar based on heading constraint filtering combined ELM

Abstract Aiming at the problem of low measurement resolution and poor tracking accuracy in High-frequency surface wave radar (HFSWR) tracking field, a tracking method that combines the heading constraint filter and Extreme Learning Machine (ELM) is proposed. Compared with the existing research, the innovation of this study lies in proposing a two-stage tracking method based on the short-term stable state of the target, and similarity of long-term trajectory. Firstly, the heading constraint information is introduced into the estimator and improve the estimation accuracy. In multi-target tracking scenarios, many tracklets can be obtained. Then, the ELM learns robust features of tracklets to achieve the trajectory segment classification from the same target. This method can be widely applied to long-term tracking of cargo or passenger ships with fixed destinations, significantly improving tracking performance by utilizing implicit domain knowledge without additional information. The actual data of this paper comes from HFSWR on Bohai Bay. Both the simulation and actual measurement experiments show that the proposed cascaded tracking method achieves long-term continuous tracking, and generates more complete trajectories. Moreover, the proposed Extended Kalman filter based on pseudo-measurement and intercept parameters (IPM-EKF) exhibits better tracking stability and accuracy in limited scan steps. This study provides new ideas and methods for improving the tracking performance of special targets in the HFSWR field by utilizing motion features.

Experimental test of optimizing maximum power point tracking performance in solar photovoltaic arrays based on backstepping control and optimized by genetic algorithm

Maximum Power Point Tracking (MPPT) is a key component of modern solar photovoltaic (PV) systems. It improves the efficiency of PV power generation systems by inevitably reducing power losses. The most developed MPPT methods adopted like hybrid approaches combine algorithms with nonlinear controllers, such as backstepping controller (BSC). Despite their prevalence, there is still a lack of standardized procedures for tuning the parameters of these nonlinear controllers. In response to this challenge, a proposed method is adopted to identify the optimal gains using Genetic Algorithm (GA) for an efficient power tracking. The advantage of this strategy is established by real-time testing with the GA-BSC technique, which validates its capacity to generate appropriate controller gains for improved tracking performance. Experimental comparisons reveal that the modified GA-BSC approach significantly outperforms the conventional GA-BSC strategy. The conventional GA-BSC method falls short due to its use of non-optimal gains, leading to less effective reference voltage tracking and lower overall performance. In contrast, the proposed method benefits from optimal gains, resulting in enhanced stability and faster response times—achieving maximum power point tracking in 0.2 s compared to 0.6 s for the P&O-BSC method. Another experiment was conducted to test the robustness of the proposed method under variable irradiation profile. The results demonstrated the efficiency of the improved GA-BSC approach compared to other methods, which was more stable and faster. These improvements highlight the capability of the proposed method to enhance hybrid method efficiency, thereby leading to increased efficiency and reliability in photovoltaic systems.

Pre-slippage detection and counter-slippage for e-pattern omniwheeled cellular conveyor

This paper presents continuation work of e-pattern omniwheeled cellular conveyor (EOCC) since its first introduction. EOCC is a conveyor that is modular and is made up of omniwheels arranged horizontally and vertically. Although in the last published paper, the EOCC had been proven to be capable of transporting box omnidirectionally and achieving yaw control concurrently, however, due to the natural properties of omniwheel, the performance is jeopardized by slippage. While minor slippage can be negligible, but a major slippage can eventually destroy the whole trajectory tracking performance. Therefore, counter-slippage methods are proposed in this paper. The simulation results show that the proposed counter-slippage method significantly improves the trajectory tracking performance up to 42% of reduction in integral of absolute error. Moreover, in this paper, pre-slippage detection method, which aims to perform early detection of slippage, is being presented as well. Although these proposed methods are simple, but they are proven to have achieved improved tracking performance than conventional controller, as presented in this paper.

A Diffusion Neural Network-Enhanced Object Tracking Approach Under Sports Scenarios

Object tracking for motion scenes is a common research concern in field of computer vision. Its goal is to accurately track targets in different time periods and predict their future states by utilizing the motion information in video sequences. However, traditional target-tracking methods in motion scenes often face challenges such as target blur, occlusion, and changes in lighting. To deal with this issue, this paper proposes a diffusion neural network-enhanced object-tracking approach under sports scenarios. In order to further improve tracking performance, the diffusion convolution operation is introduced, which propagates features at different time steps to enhance the modeling ability of target motion. Then, suitable influencing factors are selected based on motion scene object feature parameters. Finally, a target tracking method is established by integrating these two methods. In the experiment, we used a large number of real motion scene datasets to evaluate the proposed method. The experimental results show that by comparing with traditional moving object tracking methods, the proposal achieves significant improvement in tracking accuracy and methodology robustness. In addition, we also conducted stability experiments, proving that this method has good stability for models with varying kernel numbers.

Learning a Context-Aware Environmental Residual Correlation Filter via Deep Convolution Features for Visual Object Tracking

Visual tracking has become widespread in swarm robots for intelligent video surveillance, navigation, and autonomous vehicles due to the development of machine learning algorithms. Discriminative correlation filter (DCF)-based trackers have gained increasing attention owing to their efficiency. This study proposes “context-aware environmental residual correlation filter tracking via deep convolution features (CAERDCF)” to enhance the performance of the tracker under ambiguous environmental changes. The objective is to address the challenges posed by intensive environment variations that confound DCF-based trackers, resulting in undesirable tracking drift. We present a selective spatial regularizer in the DCF to suppress boundary effects and use the target’s context information to improve tracking performance. Specifically, a regularization term comprehends the environmental residual among video sequences, enhancing the filter’s discrimination and robustness in unpredictable tracking conditions. Additionally, we propose an efficient method for acquiring environmental data using the current observation without additional computation. A multi-feature integration method is also introduced to enhance the target’s presence by combining multiple metrics. We demonstrate the efficiency and feasibility of our proposed CAERDCF approach by comparing it with existing methods using the OTB2015, TempleColor128, UAV123, LASOT, and GOT10K benchmark datasets. Specifically, our method increased the precision score by 12.9% in OTB2015 and 16.1% in TempleColor128 compared to BACF.

Development of deep multi-animal tracking for domestic cats toward future application to social behavior analysis

Computer vision models such as You Only Look Once (YOLO) excel at tracking general objects, yet they often struggle to accurately track multiple animals. In this study, we aimed to verify the improvement of animal tracking performance using YOLO by fine-tuning. We obtained videos of multiple domestic cats by ourselves and annotated the videos, and performed fine-tuning with this video dataset. Results of this study show that even training with less than one hundred images improves accuracy and enables tracking of individual animals. This result suggests that our method is promising for the realization of animal social behavior analysis.

Fuzzy Logic and Neural network-based self-tuned PID controller of Quadcopter

Quadcopters have been established as a viable tool for the current era of technology. Quadcopters are also known as drones or unmanned aerial vehicles. The quadcopter can be used for aerial surveys for deliveries, filming, and missions such as detecting and mapping unreachable areas, hills, forests metropolitan areas, rocks, and obstructions that can be a danger or risk to navigation for commercial and security projects. It conducts its survey mission without human operator intervention. After the mission is accomplished, data can be fetched by GPS or the Quadcopter returns to a pre-programmed location where the data can be downloaded and processed. It can lead to an improvement in the position tracking performance due to an intrinsic change in the object's weight, buoyancy, and aerodynamic forces. In addition, quadcopters must navigate the extremely dynamic airborne environment that is represented by urban, rural, metropolitan, or natural environments. In light of this, efficient dock tracking is required to attain the requisite high precision performance, resulting in less time consumption, when the system's dynamic characteristics are time-dependent or the operating conditions and vision guidance alter. The automatic identification of objects by classifying images and self-tuning the gains of a PID (Proportional + Integral + Derivative) controller has been proposed in this paper. The object recognition is based on machine learning guided vision and self-tuning PID controller based on fuzzy logic and neural network.

Review and Analysis of RGBT Single Object Tracking Methods: A Fusion Perspective

Visual tracking is a fundamental task in computer vision with significant practical applications in various domains, including surveillance, security, robotics, and human-computer interaction. However, it may face limitations in visible light data, such as low-light environments, occlusion, and camouflage, which can significantly reduce its accuracy. To cope with these challenges, researchers have explored the potential of combining the visible and infrared modalities to improve tracking performance. By leveraging the complementary strengths of visible and infrared data, RGB-infrared fusion tracking has emerged as a promising approach to address these limitations and improve tracking accuracy in challenging scenarios. In this article, we present a review on RGB-infrared fusion tracking. Specifically, we categorize existing RGBT tracking methods into four categories based on their underlying architectures, feature representations, and fusion strategies, namely feature decoupling based method, feature selecting based method, collaborative graph tracking method, and traditional fusion method. Furthermore, we provide a critical analysis of their strengths, limitations, representative methods, and future research directions. To further demonstrate the advantages and disadvantages of these methods, we present a review of publicly available RGBT tracking datasets and analyze the main results on public datasets. Moreover, we discuss some limitations in RGBT tracking at present and provide some opportunities and future directions for RGBT visual tracking, such as dataset diversity, unsupervised and weakly supervised applications. In conclusion, our survey aims to serve as a useful resource for researchers and practitioners interested in the emerging field of RGBT tracking, and to promote further progress and innovation in this area.

Application of Gray Wolf Particle Filter Algorithm Based on Golden Section in Wireless Sensor Network Mobile Target Tracking

In order to address the issue of low tracking accuracy caused by particle depletion in the particle filter, a mobile target tracking algorithm tailored for wireless sensor networks (WSNs) is presented. This algorithm, based on the golden-section gray wolf particle filter (PF), represents a novel approach to target tracking. The algorithm’s originality lies in its ability to guide the particle swarm toward regions of higher weights, thereby striking a balance between global and local exploration capabilities. This not only alleviates issues related to sample depletion and local extrema but also enhances the diversity of the particle swarm, significantly improving tracking performance. To assess the effectiveness of the proposed algorithm, a series of simulation experiments were conducted, comparing it with the extended Kalman filter (EKF) and the standard PF algorithm. The experiments employed a constant velocity circular motion model (CM) for filtering and tracking. The root mean square error metric demonstrated a significant reduction in error of 57% and 37% in comparison to the extended Kalman filter (EKF) and the particle filter (PF), respectively. This serves to illustrate the superiority of our method in enhancing tracking accuracy.

Point-to-point virtual model control of a flexible joint robotic manipulator using trajectories of motion

This paper considers the use of trajectories of motion as a means of addressing the problem of unwanted oscillations and improving steady-state accuracy in the point-to-point motion of a flexible joint robotic manipulator using a novel third-order Poynting–Thomson virtual model control. A differentially flat model of the manipulator was exploited for classical controller design. The trajectories of motion were generated through a Newton interpolation procedure. The conditions for motion from rest to rest were derived and used to generate the trajectories required to steer the manipulator from one static point to another. Simulation and laboratory experiments confirmed that with the use of the trajectories of motion, rest-to-rest displacement is achieved without the unwanted swings about the final position. A marginal improvement in trajectory tracking performance was also verified by comparing it with the classical controller and state feedback controller. Moreover, the new controller showed better external disturbance rejection than the classical flat controller.

Uncertainty disturbance estimator control for delayed linear systems with input constraint

This paper proposes a novel Uncertainty and Disturbance Estimator (UDE) based controller for Multi-Input Multi-Output (MIMO) systems. The controller handles input constraints, time-varying matched uncertainties in state and input matrices, unknown input–output delays, and partial state measurement. The UDE error dynamics are augmented with a time-varying variable to handle input constraints effectively. The proposed controller employs a Proportional Integral Observer (PIO) and a state predictor to accurately estimate the future state of the system with only partial state measurement. We develop sufficient conditions for system stability in tracking a reference model while only an estimation of the total delay is available. Numerical simulations and experiments on a multi-rotor platform validate the proposed controller’s effectiveness, demonstrating improved tracking performance and stability compared to existing methods. The proposed UDE-based controller, with a PIO and state predictor, is capable of handling various matched uncertainties and disturbances in MIMO systems with input constraints and partial state measurement. It offers improved tracking performance and stability compared to existing studies.

Development of Augmented Reality Vision for Osteosynthesis Using a 3D Camera.

We developed a 3D camera system to track motion in a surgical field. This system has the potential to introduce augmented reality (AR) systems non-invasively, eliminating the need for the invasive AR markers conventionally required. The present study was performed to verify the real-time tracking accuracy of this system, assess the feasibility of integrating this system into the surgical workflow, and establish its potential to enhance the accuracy and efficiency of orthopedic procedures. To evaluate the accuracy of AR technology using a 3D camera, a forearm bone model was created. The forearm model was depicted using a 3D camera, and its accuracy was verified in terms of the positional relationship with a 3D bone model created from previously imaged CT data. Images of the surgical field (capturing the actual forearm) were taken and saved in nine poses by rotating the forearm from pronation to supination. The alignment of the reference points was computed at the three points of CT versus the three points of the 3D camera, yielding a 3D rotation matrix representing the positional relationship. In the original system, a stereo vision-based 3D camera, with a depth image resolution of 1280×720 pixels, 30 frames per second, and a lens field of view of 64 specifications, with a baseline of 3 cm, capable of optimally acquiring real-time 3D data at a distance of 40-60 cm from the subject was used. In the modified system, the following modifications were made to improve tracking performance: (1) color filter processing was changed from HSV to RGB, (2) positional detection accuracy was modified with supporting marker sizes of 8 mm in diameter, and (3) the detection of marker positions was stabilized by calculating the marker position for each frame. Tracking accuracy was examined with the original system and modified system for the following parameters: differences in the rotation matrix, maximum and minimum inter-reference point errors between CT-based and camera-based 3D data, and the average error for the three reference points. In the original system, the average difference in rotation matrices was 5.51±2.68 mm. Average minimum and maximum errors were 1.10±0.61 and 15.53±12.51 mm, respectively. The average error of reference points was 6.26±4.49 mm. In the modified system, the average difference in rotation matrices was 4.22±1.73 mm. Average minimum and maximum errors were 0.79±0.49 and 1.94±0.87 mm, respectively. The average error of reference points was 1.41±0.58 mm. In the original system, once tracking failed, it was difficult to recover tracking accuracy. This resulted in a large maximum error in supination positions. These issues were resolved by the modified system. Significant improvements were achieved in maximum errors and average errors using the modified system (P<0.05). AR technology using a 3D camera was developed. This system allows direct comparisons of 3D data from preoperative CT scans with 3D data acquired from the surgical field using a 3D camera. This method has the advantage of introducing AR into the surgical field without invasive markers.

Data-driven auto-tuning strategy for RTO-MPC based on Bayesian optimization

Real-time optimization (RTO) and model predictive control (MPC) are extensively employed in industrial processes to enhance economic objectives. However, the tuning of the system remains a challenge, in particular, for cases where an explicit model relating the RTO objective and the process dynamics is unknown. To address this issue, an online data-driven auto-tuning strategy leveraging two Bayesian optimization (BO) techniques is proposed. This strategy is useful for situations where the RTO objective and the process dynamics are detectable but their exact functional forms are unclear. The proposed strategy views the RTO objective as a black-box objective and interprets the steady-state conditions as black-box equality constraints within the RTO layer. In this context, the upper-trust-bound based constrained BO (UTB-CBO) method is adopted to optimize the setpoints and enhance solution feasibility. Additionally, the proposed approach can take into account measurable disturbance inputs explicitly and account for their consequential influence on objective optimization by considering disturbance variations as contextual information. Simultaneously, another contextual BO scheme is implemented to automatically tune the MPC controller parameters for improving tracking performance upon accepting the setpoints optimized by the RTO. Simulation results based on a continuous stirred-tank reactor system are given to illustrate the effectiveness of the proposed approach.

Environmental impacts of Australian pork in 2020 and 2022 determined using lifecycle assessments

Context The Australian pork industry is highly efficient, with a history of ongoing productivity and environmental improvement. The introduction of economy-wide environmental targets require delivering and tracking performance improvement. Aims This study determined carbon footprint (greenhouse gas [GHG] and land use [LU] and direct land use change [dLUC] emissions, reported as kg CO2-e), fossil energy (MJ), freshwater consumption (L), water stress (L H2O-e), land occupation (m2) and eutrophication potential (nitrogen and phosphorus) for Australian pork for 2020 and 2022. Variability between housing, manure management systems, and regions were identified, and systems analysed to determine new options for low-impact pork. Methods In the largest Australian study of its kind, data for ~70% of pigs produced were collected using a stratified design. Using attributional life cycle assessment, impacts were reported per kilogram of liveweight (LW), post-processed, retail, and boneless, fat-corrected pork. Results are presented as industry averages ± 2 × s.d. Key results Key results were 3.0 ± 0.1 and 3.0 ± 0.1 kg CO2-e GHG, 0.4 ± 0.07 and 0.3 ± 0.03 kg CO2-e LU and dLUC, 12.9 ± 0.5 and 13.4 ± 0.5 MJ, 93.8 ± 9.6 and 52.5 ± 3.6 L, 68.4 ± 6.7 and 43.2 ± 3.3 L H2O-e, and 12.0 ± 0.9 and 12.7 ± 0.9 m2/kg LW in 2020 and 2022, respectively. Due to industry growth, total emissions were higher in 2022. Eutrophication potential for Australian pork (2.2 × 10−4 ± 3.0 × 10−5 kg phosphorus and 8.7 × 10−3 ± 3.5 × 10−4 kg nitrogen/kg LW), reported for the first time, was low compared with grazing systems and European piggeries. Conclusions Industry has demonstrated long-term performance improvement, though the rate slowed between 2020 and 2022. Ongoing interventions are required to return to trend. Covered pond, deep litter, and outdoor systems produce lower carbon footprint pork and can provide other environmental benefits from renewable energy, and reduced fossil energy demand. Implications There is potential to further reduce environmental impacts through practice change. If industry is to meet formal targets, investment and proactive policy settings are required to overcome barriers to adoption of existing technology and support the techno-economic case for novel strategies.

Sensorimotor adaptation to destabilizing dynamics in weakly electric fish

Humans and other animals can readily learn to compensate for changes in the dynamics of movement. Such changes can result from an injury or changes in the weight of carried objects. These changes in dynamics can lead not only to reduced performance but also to dramatic instabilities. We evaluated the impacts of compensatory changes in control policies in relation to stability and robustness in Eigenmannia virescens, a species of weakly electric fish. We discovered that these fish retune their sensorimotor control system in response to experimentally generated destabilizing dynamics. Specifically, we used an augmented reality system to manipulate sensory feedback during an image stabilization task in which a fish maintained its position within a refuge. The augmented reality system measured the fish's movements in real time. These movements were passed through a high-pass filter and multiplied by a gain factor before being fed back to the refuge motion. We adjusted the gain factor to gradually destabilize the fish's sensorimotor loop. The fish retuned their sensorimotor control system to compensate for the experimentally induced destabilizing dynamics. This retuning was partially maintained when the augmented reality feedback was abruptly removed. The compensatory changes in sensorimotor control improved tracking performance as well as control-theoretic measures of robustness, including reduced sensitivity to disturbances and improved phase margins.

Adaptive disturbance rejection neural output feedback control of hydraulic manipulator systems

This paper proposes an adaptive disturbance rejection neural output feedback control (ADRNC) scheme for multi-degree-of-freedom (n-DOF) hydraulic manipulator systems, subjected to unknown nonlinearities, external disturbances and unmeasured system states. The controller design is formulated by integrating Radial Basis Function Neural Networks (RBFNNs) with state and disturbance observers using the backstepping method. The RBFNNs are synthesized to handle unknown nonlinear functions and the residual estimate error, coupled with external disturbances, is estimated through the combination of state observer and disturbance observer. The unique features of the proposed controller lies in its capability to estimate both matched and unmatched lumped disturbances. The auxiliary disturbance estimation law is guided by the neural learning weights and estimated system states provided by state observers. By effectively utilizing neural networks to approximate and mitigate most nonlinear uncertainties, the workload of the disturbance observer is substantially reduced. High-gain feedback is therefore avoided and improved tracking performance can be expected. Moreover, to avoid the tedious analysis and the problem of “explosion of complexity” in the conventional backstepping method, we employ a first-order sliding-mode differentiator. Rigorous analysis via Lyapunov methods establishes the stability of the entire closed-loop system, ensuring guaranteed and satisfactory tracking performance under the integrated influence of unknown nonlinearities, unmeasured states, and external disturbances. Extensive simulations are conducted to verify the effectiveness of the nested control strategy.