Motion Prediction Model Research Articles

Ground motion prediction model for shallow crustal earthquakes in Japan based on XGBoost with Bayesian optimization

Ground motion prediction is an important and complex research subject in earthquake engineering and traditional approaches based on statistical regression have much room for improvement in prediction accuracy. Utilizing the 67,164 ground motion records from KiK-net and K-Net for 777 shallow crustal earthquakes between 1997 and 2019 in Japan, this paper proposes a ground motion prediction model XGBoost-SC based on the machine learning algorithm of eXtreme Gradient Boosting (XGBoost) for Japan. Magnitude, focal depth, hypo-central distance, Vs30, site altitude, and focal mechanism were used as the feature parameters, and XGBoost, Random Forest, and Deep Neural Networks (DNN) algorithms were selected for model training while Bayesian optimization was used to search for optimized hyperparameters to improve the prediction accuracy. XGBoost algorithm was selected for further study based on the comparison of results from the three algorithms. Residual change with magnitude and hypo-central distance, the probability distribution of residuals, residual standard deviation (σ), residual mean squared error (MSE), and Pearson correlation coefficient (R) were used as the evaluation parameters, and a comparison study was performed against the ground motion prediction equation based on traditional approaches. Actual earthquake events were selected to compare the prediction results against the observation records. To further validate and explain the proposed model, SHapley Additive exPlanations (SHAP) analysis was performed to explain the impact of selected feature parameters on the proposed model. The results demonstrate that the proposed XGBoost-SC model has good prediction stability for all periods, and its residual errors are smaller than those of other models. Therefore, the proposed model can better reflect the ground motion attenuation for shallow crustal earthquakes in Japan and can serve as a better model for ground motion prediction in future aseismic design and earthquake disaster mitigation efforts.

Convolutional Neural Network for Earthquake Ground Motion Prediction Model in Earthquake Early Warning System in West Java

As urbanization continues, more people and infrastructure are concentrated in areas that are at risk from earthquakes. This can increase the potential damage and loss of life when earthquakes occur. Indonesia is a region that is near the boundary of three major tectonic plates which has a very high frequency of earthquake occurrences. Over the past two decades, a new approach to earthquake disaster risk mitigation has emerged. It is based on the advent of digital seismology and advances in data transmission and automatic processing that make it possible to send warnings before the largest ground motion that called the Earthquake Early Warning System (EEW). On-site EEW is a type of EEW that consists of limited seismic stations located at a specific destination/infrastructure (for early detection systems). On-site EEW estimates ground motion parameters directly from the characteristics of seismograms recorded by the system. An artificial intelligence approach to EEW is necessary to increase the speed and accuracy of information, which increases processing time, especially in areas very close to the epicenter

Towards explainable motion prediction using heterogeneous graph representations

Motion prediction systems play a crucial role in enabling autonomous vehicles to navigate safely and efficiently in complex traffic scenarios. Graph Neural Network (GNN)-based approaches have emerged as a promising solution for capturing interactions among dynamic agents and static objects. However, they often lack transparency, interpretability and explainability — qualities that are essential for building trust in autonomous driving systems. In this work, we address this challenge by presenting a comprehensive approach to enhance the explainability of graph-based motion prediction systems. We introduce the Explainable Heterogeneous Graph-based Policy (XHGP) model based on an heterogeneous graph representation of the traffic scene and lane-graph traversals. Distinct from other graph-based models, XHGP leverages object-level and type-level attention mechanisms to learn interaction behaviors, providing information about the importance of agents and interactions in the scene. In addition, capitalizing on XHGP’s architecture, we investigate the explanations provided by the GNNExplainer and apply counterfactual reasoning to analyze the sensitivity of the model to modifications of the input data. This includes masking scene elements, altering trajectories, and adding or removing dynamic agents. Our proposal advances towards achieving reliable and explainable motion prediction systems, addressing the concerns of users, developers and regulatory agencies alike. The insights gained from our explainability analysis contribute to a better understanding of the relationships between dynamic and static elements in traffic scenarios, facilitating the interpretation of the results, as well as the correction of possible errors in motion prediction models, and thus contributing to the development of trustworthy motion prediction systems.The code to reproduce this work is publicly available at https://github.com/sancarlim/Explainable-MP/tree/v1.1.

Development of shear-wave velocity profiles and kappa compatible with Ground Motion model for Taiwan

A new velocity model (GMM) compatible with a ground motion prediction model (GMM) for Taiwan is presented in this study. The model is developed using the data set extracted from the flatfile compiled for Taiwan Senior Seismic Hazard Analysis Committee (SSHAC) level 3 project (NCREE, 2015). The Phung et al. (2020) ergodic Sa GMM was used as a backbone model for estimating the site amplification in the frequency domain using inverse Random vibration theory (IRVT). The main motivation for developing a velocity model is that the estimation of the site scaling in terms of the time-averaged shear-wave velocity over the top 30 m (VS30), the method also provides the corresponding depth-dependent VS(z) profiles and the value for the selected VS30 value as a result of the inverse quarter-wave-length method (IQWL). Not all of the empirical amplification can be explained by the 1-D VS profiles andvalues. For soft sites (VS30 < 500 m/s), and intermediate periods (0.5-2.0 sec), there is additional amplification in the empirical data which is attributed to 3-D path and site effects. Using the proposed method, the resulting GMM provides a VS profile, k, and 3-D effect for each VS30 value, which provides a more informative handoff of ground-motion information for use in site-specific site response studies.

Learning full context feature for human motion prediction

Human motion prediction aims to predict the target poses given the previous poses. Most existing methods are devoted to extracting richer motion features from only the given previous poses to predict the target poses. However, we consider that the post poses after the target poses are helpful in acquiring the context feature and constraint between neighbor motions, which is also important for motion prediction. In this paper, we explore to make use of the post motion information for a powerful human motion prediction method. Specifically, we propose a human motion prediction model which learns the motion constraint from both the previous and post poses, in order to fully utilize the context features of the target poses. During training, the proposed memory dictionary module is used to learn the mapping from previous features to post features. In testing, the proposed memory dictionary module fully exploits the learned mapping to calculate the future motion feature according to the input previous feature. Thus, the context feature of human motion is enriched in our method. We evaluate the proposed method on two large-scale datasets, Human3.6M and CMU-Mocap. The experimental results demonstrate that our method improves the motion prediction performance, especially for long-term human motion.

Experimental comparison of linear regression and LSTM motion prediction models for MLC-tracking on an MRI-linac.

Magnetic resonance imaging (MRI)-guided radiotherapy with multileaf collimator (MLC)-tracking is a promising technique for intra-fractional motion management, achieving high dose conformality without prolonging treatment times. To improve beam-target alignment, the geometric error due to system latency should be reduced by using temporal prediction. To experimentally compare linear regression (LR) and long-short-term memory (LSTM) motion prediction models for MLC-tracking on an MRI-linac using multiple patient-derived traces with different complexities. Experiments were performed on a prototype 1.0T MRI-linac capable of MLC-tracking. A motion phantom was programmed to move a target in superior-inferior (SI) direction according to eight lung cancer patient respiratory motion traces. Target centroid positions were localized from sagittal 2D cine MRIs acquired at 4Hz using a template matching algorithm. The centroid positions were input to one of four motion prediction models. We used (1) a LSTM network which had been optimized in a previous study on patient data from another cohort (offline LSTM). We also used (2) the same LSTM model as a starting point for continuous re-optimization of its weights during the experiment based on recent motion (offline+online LSTM). Furthermore, we implemented (3) a continuously updated LR model, which was solely based on recent motion (online LR). Finally, we used (4) the last available target centroid without any changes as a baseline (no-predictor). The predictions of the models were used to shift the MLC aperture in real-time. An electronic portal imaging device (EPID) was used to visualize the target and MLC aperture during the experiments. Based on the EPID frames, the root-mean-square error (RMSE) between the target and the MLC aperture positions was used to assess the performance of the different motion predictors. Each combination of motion trace and prediction model was repeated twice to test stability, for a total of 64 experiments. The end-to-end latency of the system was measured to be (389 ± 15)ms and was successfully mitigated by both LR and LSTM models. The offline+online LSTM was found to outperform the other models for all investigated motion traces. It obtained a median RMSE over all traces of (2.8 ± 1.3)mm, compared to the (3.2 ± 1.9)mm of the offline LSTM, the (3.3 ± 1.4)mm of the online LR and the (4.4 ± 2.4)mm when using the no-predictor. According to statistical tests, differences were significant (p-value <0.05) among all models in a pair-wise comparison, but for the offline LSTM and online LR pair. The offline+online LSTM was found to be more reproducible than the offline LSTM and the online LR with a maximum deviation in RMSE between two measurements of 10%. This study represents the first experimental comparison of different prediction models for MRI-guided MLC-tracking using several patient-derived respiratory motion traces. We have shown that among the investigated models, continuously re-optimized LSTM networks are the most promising to account for the end-to-end system latency in MRI-guided radiotherapy withMLC-tracking.

Effect of Ship Motion Prediction Model on Navigational Safety

Effect of Ship Motion Prediction Model on Navigational Safety

A Human-like Inverse Kinematics Algorithm of an Upper Limb Rehabilitation Exoskeleton

Powered exoskeleton rehabilitation is an effective way to help stroke patients recover their motor abilities. Bionic structures and human-like control strategies can be used to enhance both the safety and efficacy of exoskeletons. However, the motion characteristics of the shoulder complex are not sufficiently considered. In this paper, we designed a 7-degrees-of-freedom (DOF) upper limb rehabilitation exoskeleton, FREE (functional rehabilitation exoskeleton). The mechanical structures of the shoulder and forearm of FREE are in accordance with human anatomy, and can be used to perform a wide range of synergistic motion of multiple joints while keeping a safe distance from the patient’s head. A multiple-input-multiple-output (MIMO) shoulder girdle motion prediction model was developed to satisfy the synergy between humans and exoskeletons. Moreover, a constrained task priority and projected gradient-based inverse kinematics algorithm (CTPPG-IK) was proposed to achieve assistance with scapulohumeral rhythm. A motion capture system was used to collect different activities of daily life (ADL) motion data to validate the proposed algorithm. The experimental results show that the accuracy of the prediction model is higher than that of existing models, and the inverse kinematics algorithm can handle the end-effector task and joint space with a maximum angle error of 3.04×10−3 rad.

Regional offshore ground motion prediction model from a referenced empirical approach: A case study in the Japan Trench area

With increasing resource exploitation and engineering construction offshore, the seismic safety of offshore structures has received attention from researchers and designers, and an offshore ground motion prediction model (GMPM) can be used to estimate the possible earthquake response of structures and provide the foundation for seismic fortification of structures. Because there are few strong motion records recorded by offshore stations, there are insufficient empirical data in the magnitude-distance range to directly develop a GMPM using empirical regression techniques. To develop a reliable GMPM for data-poor offshore regions, we used a referenced empirical approach based on an onshore empirical prediction model. We collected 2605 records from 138 crustal earthquake events, 2669 records from 125 slab earthquake events and 5693 records from 267 interface earthquake events recorded by S-net stations. We used the existing Japanese onshore GMPM as a reference model. Then, we analyzed the residuals of the reference model for offshore data and calculated its adjustment factors. Finally, we developed a GMPM for the Japan Trench area and compared it with the existing onshore models. The proposed model fit the data subsets at hypocentral distances of 20–300 km reasonably well. There were different spectral accelerations (SAs) for different source types, with considerably larger SAs for short periods of offshore motions from crustal events than for the motions from the other two types, and the difference becomes more obvious with the increase of magnitude. At short periods, the SAs of the offshore motions from slab events were slightly larger than those of the motions from the interface events. The discrepancy of the offshore and onshore GMPM was significant, and the peak ground accelerations (PGAs) and SAs of the offshore ground motions were apparently greater than those of onshore ground motions of the same magnitude and distance. The offshore GMPM established by the reference empirical approach contribute to studying the characteristics of offshore strong ground motions, and this approach can be applied to data-poor regions.

Sparse Bayesian Relevance Vector Machine Identification Modeling and Its Application to Ship Maneuvering Motion Prediction

In order to establish a sparse and accurate ship motion prediction model, a novel Bayesian probability prediction model based on relevance vector machine (RVM) was proposed for nonparametric modeling. The sparsity, effectiveness, and generalization of RVM were verified from two aspects: (1) the processed Sinc function dataset, and (2) the tank test dataset of the KRISO container ship (KCS) model. The KCS was taken as the main research plant, and the motion prediction models of KCS were obtained. The ε-support vector regression and υ-support vector regression were taken as the compared algorithms. The sparsity, effectiveness, and generalization of the three algorithms were analyzed. According to the trained prediction models of the three algorithms, the number of relevance vectors was compared with the number of support vectors. From the prediction results of the Sinc function and tank test datasets, the highest percentage of relevance vectors in the trained sample was below 17%. The final prediction results indicated that the proposed nonparametric models had good prediction performance. They could ensure good sparsity while ensuring high prediction accuracy. Compared with the SVR, the prediction accuracy can be improved by more than 14.04%, and the time consumption was also relatively lower. A training model with good sparsity can reduce prediction time. This is essential for the online prediction of ship motion.

Self-evaluation of automated vehicles based on physics, state-of-the-art motion prediction and user experience

Legal restrictions allow to give full control to automated vehicles for longer time periods either in restricted areas or when moving with reduced speed. Although being technically feasible for a wide range of driving scenarios, the restrictions are still in place due to the lack of a clear safety strategy. An essential step towards safety is the introduction of a self-monitoring component. In this study, a self-evaluation concept is presented which assesses a system based on a physics-defined minimum prediction horizon for state-of-the-art Deep Learning-based trajectory prediction models. Since User Experience is a key metric for car manufacturers, a further manoeuvre constraint is added to the model. We emphasize the generalizability of the presented assessment concept, however, in order to demonstrate feasibility in practical use, three specific scenarios are discussed. The results are gained with real data from publicly available driving campaigns as well as synthetically generated simulation data. Two exemplary models, a simple LSTM-based model and VectorNet, a prominent motion prediction model, are evaluated. A quantitative assessment shows a lack of training data in the public datasets for vehicle speeds > 25 m/s in order to offer safe driving above such vehicle speeds.

Task-Constrained Motion Planning Considering Uncertainty-Informed Human Motion Prediction for Human–Robot Collaborative Disassembly

While the disassembly of high-precision electronic devices is a predominantly labor-intensive process, collaborative robots provide a promising solution through human–robot collaboration. To ensure efficient yet safe collaboration, this article presents a new way to generate task-constrained and collision-free motion for a collaborative robot operating in a dynamic environment involving human movement, which is traditionally challenging due to the high degree of freedom of the corobot and the uncertainty nature of human motion. We first establish a neural human motion prediction model with quantified uncertainty, and then optimize the configuration of the robot online by taking the human motion and uncertainties into consideration. While such rationale is straightforward in nature, our method explicitly quantified the uncertainty of the neural human prediction model to further enhance the collaboration safety, and integrated the quantified uncertainty into the task-satisfied motion planning in real time to efficiently conduct tasks. Extensive experimental tests and comparison studies have been conducted to validate the efficiency and effectiveness of the proposed planning method.

Towards trustworthy multi‐modal motion prediction: Holistic evaluation and interpretability of outputs

AbstractPredicting the motion of other road agents enables autonomous vehicles to perform safe and efficient path planning. This task is very complex, as the behaviour of road agents depends on many factors and the number of possible future trajectories can be considerable (multi‐modal). Most prior approaches proposed to address multi‐modal motion prediction are based on complex machine learning systems that have limited interpretability. Moreover, the metrics used in current benchmarks do not evaluate all aspects of the problem, such as the diversity and admissibility of the output. The authors aim to advance towards the design of trustworthy motion prediction systems, based on some of the requirements for the design of Trustworthy Artificial Intelligence. The focus is on evaluation criteria, robustness, and interpretability of outputs. First, the evaluation metrics are comprehensively analysed, the main gaps of current benchmarks are identified, and a new holistic evaluation framework is proposed. Then, a method for the assessment of spatial and temporal robustness is introduced by simulating noise in the perception system. To enhance the interpretability of the outputs and generate more balanced results in the proposed evaluation framework, an intent prediction layer that can be attached to multi‐modal motion prediction models is proposed. The effectiveness of this approach is assessed through a survey that explores different elements in the visualisation of the multi‐modal trajectories and intentions. The proposed approach and findings make a significant contribution to the development of trustworthy motion prediction systems for autonomous vehicles, advancing the field towards greater safety and reliability.

Ground Motion Prediction of High-Energy Mining Seismic Events: A Bootstrap Approach

Induced seismicity has been a serious problem for many coal mines in the Upper Silesian Coal Basin in Poland for many decades. The occurring mining tremors of the rock mass generate seismic vibrations that cause concern to the local population and in some rare cases lead to partial damage to buildings on the surface. The estimation of peak ground acceleration values caused by high energy mining seismic tremors is an important part of seismic hazard assessment in mining areas. A specially designed bootstrapping procedure has been applied to estimate the ground motion prediction model and makes it possible to calculate the confidence intervals of these peak ground acceleration values with no assumptions about the statistical distribution of the recorded seismic data. Monte Carlo sampling with the replacement for 132 seismic records measured for mining seismic tremors exceeding 150 mm/s2 have been performed to estimate the mean peak ground acceleration values and the corresponding upper limits of 95% confidence intervals. The specially designed bootstrap procedure and obtained ground motion prediction model reflect much better the observed PGA values and therefore provide more accurate PGA estimators compared to the GMPE model from multiple regression analysis. The bootstrap analysis of recorded peak ground acceleration values of high-energy mining tremors provides significant information on the level of seismic hazard on the surface infrastructure. A new tool has been proposed that allows for more reliable determination of PGA estimators and identification in the areas in coal mines that are prone to high-energy seismic activity.

Evaluating motion of pancreatic tumors and anatomical surrogates using cine MRI in 0.35T MRgRT under free breathing conditions.

Treatment tolerability is a significant limitation to pancreatic cancer treatment with radiotherapy due to proximity to highly radiosensitive organs and respiratory motion necessitating expanded target margins. Further, pancreatic tumors are difficult to visualize on conventional radiotherapy systems. Surrogates are often used to locate the tumor but are often inconsistent and do not provide strong positional relations throughout the respiratory cycle. This work utilizes a retrospective dataset of 45 pancreatic cancer patients treated on an MR-Linac system with cine MRI acquired for real-time target tracking. We investigated intra-fraction motion of tumors and two abdominal surrogates, leading to prediction models between the tumor and surrogate. Patient specific motion evaluation and prediction models were generated from 225 cine MRI series acquired during treatment. Tumor contours were used to evaluate the pancreatic tumor motion. Linear regression and principal component analysis (PCA) based models were used to predict tumor position from the anterior-posterior (AP) motion of the abdominal surface, the superior-inferior (SI) motion of the diaphragm, or a combination. Models were evaluated using mean squared error (MSE) and mean absolute error (MAE). Contour analysis showed the average pancreatic tumor motion range was 7.4±2.7mm and 14.9±5.8mm in the AP and SI directions, respectively. The PCA model had MSE of 1.4mm2 and 0.6mm2 , for the SI and AP directions, respectively, with both surrogates as inputs for the models. When only the abdomen surrogate was used, MSE was 1.3mm2 and 0.4mm2 in the SI and AP directions, while it was 0.4mm2 and 1.3mm2 when only the diaphragm surrogate was used. We evaluated intra-fraction pancreatic tumor motion and demonstrated prediction models between the tumor and surrogate. The models calculated the pancreatic tumor position from diaphragm, abdominal, or both contours within standard pancreatic cancer target margin, and the process could be applied to other disease sites in the abdominothoracic cavity.

Characteristic of hydrodynamic derivatives for a multi-hull navigation vessel in dynamic positioning mode

The Quadmaran vessel is a maritime multi-hull vessel developed for water quality observation in aquaculture farms. The vessel’s dynamic positioning (DP) performance is essential for the accuracy of the collected data. The maneuvering characteristics in DP mode have been investigated in our previous research using a quasi-static motion prediction model based on static planar motion mechanism (PMM) tests. However, since the Quadmaran vessel has a unique hull arrangement, the motion characteristics need to be determined with greater accuracy. Therefore, in this study, the maneuvering equations are newly established with higher accuracy hydrodynamic derivatives determined through dynamic PMM tests. The higher order hydrodynamic derivatives are obtained from pure sway, pure yaw and pure yaw with a constant angle tests. Model scale experiments are conducted to establish an advanced mathematical model and compared with the quasi-static assumption results. The result shows significant improvements in the motion prediction under DP mode, and a more accurate motion prediction model is successfully established. This study is a new attempt to utilize the dynamic PMM tests method on a vessel with unique shape and motion characteristics. The advanced mathematical model deepens our understanding of the maneuverability of a multi-hull vessel in DP mode.

Weighted multi-kernel relevance vector machine for 3 DOF ship manoeuvring modeling with full-scale trial data

In order to improve the accuracy of ship maneuvering state prediction, a nonparametric identification method based on weighted multi-kernel relevance vector machine (WMRVM) is proposed to establish the motion prediction models of the vessel YUKUN. The WMRVM is composed of weight parameters and different types of kernel functions. On the one hand, RVM provides probability interpretation, its kernel functions do not have to satisfy Mercer condition; on the other hand, WMRVM improves the prediction accuracy of single kernel RVM. Taking the vessel YUKUN as the research plant, the full-scale trials satisfy the requirements of maneuverability test. The 20°/20° zigzag test data and 30° turning test data are used as the generalization verification. The prediction results obtained by the proposed algorithm are compared with those obtained by RVM and support vector regression (ε-SVR). The results show that the proposed algorithm has higher prediction accuracy and better generalization. This research can lay the foundation for the application of RVM in ship maneuvering state prediction and online identification modeling.

Model Predictive Interaction Control for Robotic Manipulation Tasks

This article presents the concept of model predictive interaction control (MPIC) as a generic, flexible, and comprehensive approach for robotic manipulation tasks. MPIC is based on the repetitive solution of an optimal control problem that includes a robot model for motion prediction as well as an interaction model for force prediction. In order to handle both elastic and rigid contact situations, a cascaded approach with low-level PD control is adopted, which allows to combine the linear-elastic environment model and the limited controller stiffness. Due to its flexibility, MPIC can be favorably used for realizing the elementary manipulation primitives (MP) within a hierarchical task planning framework, where each MP corresponds to a particular parameterization of the cost function and the constraints. The control methodology and the manipulation approach are evaluated in simulations and experiments using a 7-degree-of-freedom industrial robot.

A far-field ground motion prediction model for interface earthquakes at the hill zone of Mexico City

This study presents an updated attenuation model to predict the peak ground acceleration (PGA), peak ground velocity (PGV), 5% damped pseudo-spectral acceleration (SA), and the average spectral acceleration (AvgSA) at the hill zone of Mexico City for interface earthquakes. The strong-motion dataset comprises 33 earthquakes recorded at CU station, covering a moment magnitude (Mw) range from 6.0 to 8.1 and a source-to-site distance (Rrup) range from 240 to 490 km. Given the small number of available observations, a Bayesian regression scheme is used to obtain the coefficients of the ground-motion prediction model (GMPM). In addition, the epistemic uncertainty in the estimation of the regression coefficients is evaluated, showing its impact on the framework of a probabilistic seismic hazard analysis (PSHA). The results are compared with models previously developed for the CU station, discussing the differences observed between the median predictions and their standard deviations. Likewise, seismic hazard curves are computed and compared with empirical curves obtained by counting the number of times per year that a given value of ground-motion intensity is exceeded. The results show that the dispersion of the GMPM proposed is lower than the previous models for PGA and SA, which means better predictability and more reliable estimates of the seismic hazard at the site.

Message from the Winner

We are very honored to receive the JDR Award for the Most Downloaded Article 2022, a new award from the Journal of Disaster Research (JDR). After the 2011 Great Tohoku Earthquake, it is now said that a magnitude 9-class earthquake could occur in the Nankai Trough, and it is necessary for us to know how strong ground motion (shaking) we will experience during such a great earthquake in order to take countermeasures. In this paper, we developed a new seismic motion prediction model that is applicable to M9-class earthquakes. We did this by adding the strong-motion records of the 2011 Great Tohoku Earthquake to the many records accumulated by various organizations since the 1950s. We would like to express our gratitude to all the organizations that provided us with their valuable records. Nearly 10 years have passed since the publication of the paper, and the records accumulated during that time have helped us to recognize the usefulness of our proposed model. We understand that our paper is still being downloaded in large numbers today and that our model is now being used not only in research but also in practice. As natural disasters occur more frequently in Japan and around the world, we hope that JDR will continue to grow and develop as an international source of useful knowledge on natural disasters.