Motion Model Research Articles

CvTMorph: Improving Local Feature Extraction in Medical Image Registration for Respiratory Motion Modeling with Convolutional Vision Transformer.

Accurately modeling respiratory motion in medical images is crucial for various applications, including radiation therapy planning. However, existing registration methods often struggle to extract local features effectively, limiting their performance. In this paper, we aimed to propose a new framework called CvTMorph, which utilizes a Convolutional vision Transformer (CvT) and Convolutional Neural Networks (CNN) to improve local feature extraction. CvTMorph integrates CvT and CNN to construct a hybrid model that combines the strengths of both approaches. Additionally, scaling and square layers are added to enhance the registration performance. We have evaluated the performance of CvTMorph on the 4D-Lung and DIR-Lab datasets and compared it with state-of-the-art methods to demonstrate its effectiveness. The experimental results have demonstrated CvTMorph to outperform the existing methods in terms of accuracy and robustness for respiratory motion modeling in 4D images. The incorporation of the convolutional vision transformer has significantly improved the registration performance and enhanced the representation of local structures. CvTMorph offers a promising solution for accurately modeling respiratory motion in 4D medical images. The hybrid model, leveraging convolutional vision transformer and convolutional neural networks, has proven effective in extracting local features and improving registration performance. The results have highlighted the potential of CvTMorph for various applications, such as radiation therapy planning, and provided a basis for further research in this field.

Estimation of the error in determining the motion parameters of maneuvering objects

The study contains the definition of errors in estimating the motion parameters of maneuvering objects that arise due to the discrepancy between the real model and the model used to build the filter. In the real model, along with noise and stochastic uncertainties, non-stochastic uncertainties act. The solution to the problem of estimation errors is based on the use of convex polyhedra. The statement of the research problem is given. It is shown that since the observer uses the motion model and the measurement equation, the source of estimation errors is not only the above, but also the fact that along with the random component of the measurement noise, there is also a non-stochastic component that is not taken into account by the observer. At the same time, the statistical characteristics of the latter do not exist or are unknown. It is shown that the use of a filter, for example, Kalman, will lead to errors. The conditions that determine the average filtering error and the total standard error of the estimation are described. It is noted that the obtained calculation results allow us to estimate the maximum permissible values used in the design of the filter for the most unfavorable case of evaluation. The relationship between the filtration quality and the parameters of the effects is determined. Expressions are obtained regarding the filtration quality for the worst and the best for object maintenance. The asymptotic properties of filtering errors are considered. It is shown that the "convergence" of the filter means its final "memory", since errors at the next step do not depend on errors from all previous steps, but only on steps from a certain value. The conditions under which the potential accuracy of the filter is achieved are also determined. The solution of extreme problems is described. You can determine the maximum filtering quality by iterating over the values on a set of corner points or using extremum conditions. Numerical simulation of extreme values of filtering errors is given. The asymptotic properties of filtering errors are considered. It is shown that the "convergence" of the filter means its final "memory", since errors at the next step do not depend on errors from all previous steps, but only on steps from a certain value. The conditions under which the potential accuracy of the filter is achieved are also determined. The solution of extreme problems is described. You can determine the maximum filtering quality by iterating over the values on a set of corner points or using extremum conditions. Numerical simulation of extreme values of filtering errors is given. Expressions for the set of average values of filtering errors are found, asymptotic and extreme properties of errors and sets are considered. It is shown that the conducted analysis of the accuracy of the filtering algorithm is evaluative in nature and is designed for the worst case. If the accuracy is not satisfactory, it is necessary to either change the model or synthesize a minimax-stochastic filter.

Drivers to seismic hazard curve slope

AbstractThe slope of a linear approximation of a probabilistic seismic hazard curve, when it is represented in the log‐log scale, is a key parameter for seismic risk assessment based on closed‐form solutions, and other applications. On the other hand, it is observed that different hazard models can provide, at the same site, comparable ground shaking, yet appreciably different slopes for the same exceedance return period. Moreover, the slope at a given return period can increase or decrease from low‐ to high‐hazardous sites, depending on the models the probabilistic seismic hazard analysis (PSHA) is based on. In the study, the sensitivity of the slope to the main model components involved in PSHA was explored, that is: the earthquake rate, the magnitude and source‐to‐site distance distributions, and the value of the residual of ground motion models (GMM). With reference to a generic site, affected by an ideal seismic source zone, where magnitude follows the Gutenberg‐Richter (G‐R) relationship, it was found that the local slope of hazard curve increases with the following factors in descending order of importance: (i) increasing distance from the source; (ii) decreasing maximum magnitude and increasing ‐value of the G‐R model; (iii) increasing rate of earthquakes of interest; (iv) increasing residual of the GMM. These results help explain the systematic differences in hazard curve slopes found in three authoritative hazard models for Italy, and the related impact on simplified risk assessment.

KISS-Keep It Static SLAMMOT-The Cost of Integrating Moving Object Tracking into an EKF-SLAM Algorithm.

The treatment of moving objects in simultaneous localization and mapping (SLAM) is a key challenge in contemporary robotics. In this paper, we propose an extension of the EKF-SLAM algorithm that incorporates moving objects into the estimation process, which we term KISS. We have extended the robotic vision toolbox to analyze the influence of moving objects in simulations. Two linear and one nonlinear motion models are used to represent the moving objects. The observation model remains the same for all objects. The proposed model is evaluated against an implementation of the state-of-the-art formulation for moving object tracking, DATMO. We investigate increasing numbers of static landmarks and dynamic objects to demonstrate the impact on the algorithm and compare it with cases where a moving object is mistakenly integrated as a static landmark (false negative) and a static landmark as a moving object (false positive). In practice, distances to dynamic objects are important, and we propose the safety-distance-error metric to evaluate the difference between the true and estimated distances to a dynamic object. The results show that false positives have a negligible impact on map distortion and ATE with increasing static landmarks, while false negatives significantly distort maps and degrade performance metrics. Explicitly modeling dynamic objects not only performs comparably in terms of map distortion and ATE but also enables more accurate tracking of dynamic objects with a lower safety-distance-error than DATMO. We recommend that researchers model objects with uncertain motion using a simple constant position model, hence we name our contribution Keep it Static SLAMMOT. We hope this work will provide valuable data points and insights for future research into integrating moving objects into SLAM algorithms.

Towards Autonomous Operation of UAVs Using Data-Driven Target Tracking and Dynamic, Distributed Path Planning Methods

A hybrid real-time path planning method has been developed that employs data-driven target UAV trajectory tracking methods. It aims to autonomously manage the distributed operation of multiple UAVs in dynamically changing environments. The target tracking methods include a Gaussian mixture model, a long short-term memory network, and extended Kalman filters with pre-specified motion models. Real-time vehicle-to-vehicle communication is assumed through a cloud-based system, enabling virtual, dynamic local networks to facilitate the high demand of vehicles in airspace. The method generates optimal paths by adaptively employing the dynamic A* algorithm and the artificial potential field method, with minimum snap trajectory smoothing to enhance path trackability during real flights. For validation, software-in-the-loop testing is performed in a dynamic environment composed of multiple quadrotors. The results demonstrate the framework’s ability to generate real-time, collision-free flight paths at low computational costs.

Generalized interacting multiple model Kalman filtering algorithm for maneuvering target tracking under non-Gaussian noises

The traditional interacting multiple model Kalman filtering algorithm (IMM-KF) can deal with the maneuvering target problem under Gaussian noise by soft switching among possible motion models. In practice, its performance is likely to degrade when handling non-Gaussian noise. We introduce the Gaussian mixture model (GMM) into the IMM-KF, and the GMM is utilized to model the non-Gaussian measurement noise as a mixture of multiple Gaussian probability densities with a certain probability. Then, a GIMM framework is proposed that enables accurate switching and fusion among multiple possible motion and noise models. And combined with Kalman filtering (KF), a GIMM-KF algorithm is proposed that enables accurate state estimation of maneuvering targets under non-Gaussian noise conditions. Subsequently, the provided simulations and experiments validate that the GIMM-KF algorithm outperforms existing methods in terms of accuracy, stability and robustness.

Directional self-migration of droplets on an inclined surface driven by wettability gradient

In the current study, the anti-gravity directional self-migration of droplets on an inclined surface driven by wettability gradient (ω) was investigated using a front-tracking method. A unified mechanical model of droplet motion on an inclined wettability gradient wall was derived, considering the driving force generated by ω (Fd), gravity (G), and flow resistance (Ff). The model demonstrates that ω, G, and inclination angle (α) are key parameters affecting droplet motion. By varying ω, Bond number (Bo), and α, the droplet dynamic characteristics were analyzed, and a real-time Capillary number (Ca) was introduced to measure the droplet migration speed. The results indicate that a larger ω generates a greater Fd, leading to faster migration and more pronounced spreading. When the ratio of the channel width to the droplet diameter is 0.7, the droplet can cross three regions, obtaining double Fd, and Ca curve exhibits a bimodal structure. When the ratio of the channel width to the droplet diameter is 1.2, the droplet slides and spreads in the middle region without ω, resulting in a trimodal Ca curve. A larger Bo implies a stronger gravity effect, reducing the net driving force for upward migration and slowing the migration speed. At α=30° and ω=0.54, Bo reaches its critical value at 0.5, where G exceeds Fd, causing the droplet to slide downward along the wall. α affects droplet motion by controlling the gravitational component along the wall (Gx). A larger α results in a smaller net driving force for upward migration, reducing the migration speed.

Analytic optimal control for multi-satellite assembly using linearized twistor-based model

This paper presents Guidance and Control (G&C) systems for multi-satellite assembly in proximity operations. The systems utilize the twistor model, which is linearized through Taylor’s series. Decentralized control laws, designed using Linear Quadratic Regulator (LQR) and Model Predictive Control (MPC), are employed to track an energy-optimal trajectory generated using the Hamiltonian approach. Data exchange between satellites and their neighbors is represented using graph theory. The decentralized MPC framework is implemented using the CasADi package. To ensure collision avoidance between the satellites, a repulsive control law is designed, considering symmetric input saturation in the actuators. The proposed G&C systems are tested using a high-fidelity nonlinear satellite relative motion model that incorporates orbital perturbations. Numerical simulations are performed in a MATLAB® environment, and the results are visualized using STK®. Furthermore, a comparative study is conducted to evaluate tracking performance and fuel consumption between the two control methods. The results demonstrate that the use of an optimal trajectory reduces fuel consumption for both control algorithms.

NEFELI: A deep-learning detection and tracking pipeline for enhancing autonomy in advanced air mobility

Efficient detection and accurate collision estimation for non-cooperative aerial vehicles are crucial for the realization of fully autonomous aircraft and Advanced Air Mobility (AAM). This paper introduces NEFELI, a machine learning software utilizing optical sensors to detect and track non-cooperative aerial vehicles. NEFELI's detector employs an enhanced YOLOv5-large model, strengthened with a sliced inference step to enhance detection capabilities for distant, diminutive objects. Furthermore, NEFELI introduces several innovations in its tracking component. A key advancement lies in the creation and utilization of the first large-scale re-identification (Re-ID) dataset of aerial objects. This dataset is used to train the deep learning appearance (Re-ID) model of the tracking module and integrates appearance information into the detection and tracking pipeline, resulting in more robust and reliable tracking performance. Moreover, the tracking model combines the deep learning appearance model with a Kalman Filter-based motion model to address the challenge of precisely tracking distant aerial objects. Notably, an extensive comparative analysis that was conducted showed that NEFELI outperforms state-of-the-art detection and tracking models in terms of Higher Order Tracking Accuracy (HOTA) metric, ID switches, and Association Accuracy (AssA) by a wide margin. A crucial aspect of this work is NEFELI's software architecture design, which enables efficient implementation on a low SWaP (Size, Weight, and Power) edge Graphic Processing Unit (GPU). To further showcase NEFELI's generalization capabilities and edge implementation performance, real-world flight experiments with small UAVs were carried out. The experimental results demonstrate NEFELI's ability to detect and track small UAVs at distances of up to 145 m in real-time speed of 6.7 fps.

Computational account for the naturalness perception of others' jumping motion based on a vertical projectile motion model.

The visual naturalness of a rendered character's motion is an important factor in computer graphics work, and the rendering of jumping motions is no exception to this. However, the computational mechanism that underlies the observer's judgement of the naturalness of a jumping motion has not yet been fully elucidated. We hypothesized that observers would perceive a jumping motion as more natural when the jump trajectory was consistent with the trajectory of a vertical projectile motion based on Earth's gravity. We asked human participants to evaluate the naturalness of point-light jumping motions whose height and duration were modulated. The results showed that the observers' naturalness rating varied with the modulation ratios of the jump height and duration. Interestingly, the ratings were high even when the height and duration differed from the actual jump. To explain this tendency, we constructed computational models that predicted the theoretical trajectory of a jump based on the projectile motion formula and calculated the errors between the theoretical and observed trajectories. The pattern of the errors correlated closely with the participants' ratings. Our results suggest that observers judge the naturalness of observed jumping motion based on the error between observed and predicted jump trajectories.

Disturbance Estimation and Predefined-Time Control Approach to Formation of Multi-Spacecraft Systems.

Accurate sensing and control are important for high-performance formation control of spacecraft systems. This paper presents a strategy of disturbance estimation and distributed predefined-time control for the formation of multi-spacecraft systems with uncertainties based on a disturbance observer. The process begins by formulating a kinematics model for the relative motion of spacecraft, with the formation's communication topology represented by a directed graph for the formation system of the spacecraft. A disturbance observer is then developed to estimate the disturbances, and the estimation errors can be convergent in fixed time. Following this, a disturbance-estimation-based sliding mode control is proposed to guarantee the predefined-time convergence of the multi-spacecraft formation system, regardless of initial conditions. It allows each spacecraft to reach its desired position within a set time frame. The results of the analysis of the multi-spacecraft formation system are also provided. Finally, an example simulation of a five-spacecraft formation flying system is provided to demonstrate the presented formation control method.

Earth’s tectonic and plate boundary evolution over 1.8 billion years

Understanding the intricate relationships between the solid Earth and its surface systems in deep time necessitates comprehensive full-plate tectonic reconstructions that include evolving plate boundaries and oceanic plates. In particular, a tectonic reconstruction that spans multiple supercontinent cycles is important to understand the long-term evolution of Earth’s interior, surface environments and mineral resources. Here, we present a new full-plate tectonic reconstruction from 1.8 Ga to present that combines and refines three published models: one full-plate tectonic model spanning 1 Ga to present and two continental-drift models focused on the late Paleoproterozoic to Mesoproterozoic eras. Our model is constrained by geological and geophysical data, and presented as a relative plate motion model in a paleomagnetic reference frame. The model encompasses three supercontinents, Nuna (Columbia), Rodinia, and Gondwana/Pangea, and more than two complete supercontinent cycles, covering ∼40% of the Earth’s history. Our refinements to the base models are focused on times before 1.0 Ga, with minor changes for the Neoproterozoic. For times between 1.8 Ga and 1.0 Ga, the root mean square speeds for all plates generally range between 4 cm/yr and 7 cm/yr (despite short-term fast motion around 1.1 Ga), which are kinematically consistent with post-Pangean plate tectonic constraints. The time span of the existence of Nuna is updated to between 1.6 Ga (1.65 Ga in the base model) and 1.46 Ga based on geological and paleomagnetic data. We follow the base models to leave Amazonia/West Africa separate from Nuna (as well as Western Australia, which only collides with the remnants of Nuna after initial break-up), and South China/India separate from Rodinia. Contrary to the concept of a “boring billion”, our model reveals a dynamic geological history between 1.8 Ga and 0.8 Ga, characterized by supercontinent assembly and breakup, and continuous accretion events. The model is publicly accessible, providing a framework for future refinements and facilitating deep time studies of Earth’s system. We suggest that the model can serve as a valuable working hypothesis, laying the groundwork for future hypothesis testing.

Immersion and Invariance-Based Linear Tracking and Regulation Controller for Depth Position of an AUV

This paper addresses the tracking control challenge in the diving motion system of a specific class of autonomous underwater vehicles (AUVs) characterized by a torpedo-like shape. A decoupled and reduced-order three degrees-of-freedom linearized diving motion model is employed for depth position control. A control law is synthesized using the immersion and invariance (I&I) technique to achieve the control objectives. The primary aim is to attain tracking by immersing a stable, lower-order target (second-order) dynamic system into a three-dimensional manifold, upon which the closed-loop system evolves. We address the regulation problem as a specialized instance of the tracking problem, with the reference input set as a predetermined known depth that requires regulation. The efficacy of the proposed control law is evaluated through simulation studies involving various scenarios. Robustness tests are conducted to assess the control law’s performance under modeling uncertainties and underwater disturbances. The computer simulation employs an AUV named MAYA, utilizing experimentally validated diving motion parameters. A comparative analysis is performed between the proposed control law and other benchmark controllers to gauge its performance. Additionally, the effectiveness of the proposed control law is confirmed by validating its application to the nonlinear model of the diving motion system.

ITRF2020 seasonal geocenter motion model

ITRF2020 seasonal geocenter motion model

Probabilistic solution of non-linear random ship roll motion by data-driven method

In this paper, a data-driven method is employed to investigate the probability density function (PDF) of nonlinear stochastic ship roll motion. The mathematical model of ship roll motion comprises a linear term with cubic damping and a nonlinear restoring moment represented as an odd-degree polynomial up to the fifth order. The data-driven method integrates maximum entropy, the pseudo-inverse algorithm, and a backpropagation (BP) neural network to obtain the PDF. The process begins with simulating data for the nonlinear stochastic system, followed by dimensional analysis to identify dimensionless parameter clusters. Optimization algorithms are then employed to solve for the coefficients, leading to the development of a BP neural network model trained to predict the PDF across various system characteristics and excitation intensities. The method's effectiveness is validated with Monte Carlo simulations, demonstrating high accuracy and reduced sensitivity to parameter variations.

Quaternion Regularization of Differential Equations of Perturbed Central Motion and Regular Models of Orbital (Trajectory) Motion: Review and Analysis of Models, Their Applications

Quaternion Regularization of Differential Equations of Perturbed Central Motion and Regular Models of Orbital (Trajectory) Motion: Review and Analysis of Models, Their Applications

Evaluation of a Coupled CFD and Multi-Body Motion Model for Ice-Structure Interaction Simulation

The interaction of water flow, ice, and structures is common in fluvial ice processes, particularly around Ice Control Structures (ICSs) that are used to manage and prevent ice jam floods. To evaluate the effectiveness of ICSs, it is essential to understand the complex interaction between water flow, ice and the structure. Numerical modeling is a valuable tool that can facilitate such understanding. Until now, classical Eulerian mesh-based methods have not been evaluated for the simulation of ice interaction with ICS. In this paper we evaluate the capability, accuracy, and efficiency of a coupled Computational Fluid Dynamic (CFD) and multi-body motion numerical model, based on the mesh-based FLOW-3D V.2023 R1 software for simulation of ice-structure interactions in several benchmark cases. The model’s performance was compared with results from meshless-based models (performed by others) for the same laboratory test cases that were used as a reference for the comparison. To this end, simulation results from a range of dam break laboratory experiments were analyzed, encompassing varying numbers of floating objects with distinct characteristics, both in the presence and absence of ICS, and under different downstream water levels. The results show that the overall accuracy of the FLOW-3D model under various experimental conditions resulted in a RMSE of 0.0534 as opposed to an overall RMSE of 0.0599 for the meshless methods. Instabilities were observed in the FLOW-3D model for more complex phenomena that involve open boundaries and a larger number of blocks. Although the FLOW-3D model exhibited a similar computational time to the GPU-accelerated meshless-based models, constraints on the processors speed and the number of cores available for use by the processors could limit the computational time.

Distributed control of spacecraft formation under [formula omitted] perturbation in the port-Hamiltonian framework

To control the relative position and relative velocity of spacecraft formation, a time-varying nonlinear relative motion model under J2 perturbation in an elliptical orbit is established in the port-Hamiltonian (PH) framework, and a distributed control law for spacecraft formation is developed using the consensus algorithm for parameter estimation and the passivity-based control (PBC) method based on the state-error interconnection and damping assignment (IDA) technique. First, the influence of J2 perturbation on the potential energy and orbit parameters in the model is considered when the relative motion model is built in the PH frame, and the expression of the relative motion acceleration under J2 perturbation is given. Second, to solve the problem that the state of the chief spacecraft cannot be directly obtained from the deputy spacecraft under distributed communication, a consensus-based parameter estimation method is introduced, the estimated parameter of the chief spacecraft is applied to the relative motion model in the PH frame, and a state error model with the estimated parameters derived from the consistency algorithm is established. Then, after the stability analysis is conducted on the desired PH system containing the estimated parameter values, the desired Hamiltonian energy function with the estimated parameters is designed according to the time-varying errors of the relative equilibrium states of the deputy spacecraft and the errors between deputy spacecraft, and the distributed control law of spacecraft formation is derived based on the state-error IDA-PBC method. Finally, the expected relative motion trajectory designed based on the TH equation is used to simulate a spacecraft formation under J2 perturbation, and the results indicate that the spacecraft can quickly converge to the expected formation under the influence of the control law.

A quickly discrete-continuous phase interaction model for pore-scale motion of droplets in porous foams

The expensive computational cost of pore-scale modeling of droplets in porous foams limits the understanding and application of spray cooling techniques. Thus, a numerical model that can cheaply and quickly predict the interaction of moving droplets with the foam skeleton is valuable. In this work, a novel model based on foam geometry and octree (FGAO) is proposed for the rapid evaluation of discrete-continuous phase interactions in porous foams. The performance of the FGAO model is investigated for different Stokes and Weber numbers, with the full pore-scale model as a benchmark. The results show that the accuracy improves as the Stokes number increases, but the effect of the Weber number on the accuracy is complicated. The accuracy decreases as the We close to the intersection of the interaction region, while it improves as the Weber number is far away from the intersection of the interaction region. To quantitatively assess the error, a relative error (RE) and a weighted mean absolute percentage error (WMAPE) are introduced. The results show that the FGAO model has excellent accuracy with RE < 2.79 % at Stk > 10. At the criteria of WMAPE, the FGAO model performs well at the condition of the initial Weber number is far away from the intersection of the region of the Bai-Gosman model, with WMAPE < 7.3 % at We = 1 Stk = 50 and WMAPE < 3.9 % at We = 400 Stk = 50. Finally, the current method offers a significant advantage in computational speed compared to the full pore-scale model, with an improvement of about 62.9 times for Stk = 50, We = 10.

Dynamic nonparametric modeling of sail-assisted ship maneuvering motion based on GWO-KELM

Accurately predicting the state of ship motion is crucial for ship navigation control with energy efficiency optimization studies. When modeling the motion of novel ships equipped with propulsion devices such as sails, it is essential to consider the effects of sail thrust and lateral forces on the ship's motion states. Building upon the analysis of sail force and sail-assisted ship motion modeling, this paper proposes a ship motion dynamic modeling method utilizing Kernel Extreme Learning Machine (KELM) regression. By optimizing the hyperparameters of the kernel function using the Grey Wolf Optimizer (GWO) algorithm and incorporating a sliding time window for dynamic model updating, the proposed approach combines the efficient learning mechanism of KELM with the global search capability of GWO, effectively enhancing prediction accuracy and timeliness. Utilizing KVLCC2 ship model motion data, simulation data from the sail-assisted ship “New Aden” and real ship data as case studies, comparative analysis with traditional models such as SVM and ELM. The results demonstrate that the GWO-KELM modeling approach exhibits robustness and high prediction accuracy, the MAE of the proposed model in the real ship motion prediction effect is reduced by 8.89% and the RMSE is reduced by 9.44% compared with the other models, which significantly enhance the predictive performance of sail-assisted ship maneuvering motion states.