Motion Prediction Research Articles

Nonparametric Modeling and Control of Ship Steering Motion Based on Local Gaussian Process Regression

This paper aims to study the nonparametric modeling and control of ship steering motion. Firstly, the black box response model is derived based on the Nomoto model. Then, the establishment of a nonparametric response model and prediction of ship steering motion are realized by applying the local Gaussian process regression (LGPR) algorithm. To assess the performance of LGPR, two cases are studied, including a Mariner class vessel by using simulation data and a KVLCC2 tanker model by using experimental data. The results reveal that the response model identified by LGPR presents good prediction accuracy and low computational burden. Finally, the identified response model is used as the basis for developing the ship heading controller, and the results demonstrate that the proposed controller is able to achieve good dynamic performance.

3D Printed Conformal Strain and Humidity Sensors for Human Motion Prediction and Health Monitoring via Machine Learning.

Wearable sensors have garnered considerable attention due to their flexibility and lightweight characteristics in the realm of healthcare applications. However, developing robust wearable sensors with facile fabrication and good conformity remains a challenge. In this study, a conductive graphene nanoplate-carbon nanotube (GC) inkis synthesized for multi jet fusion (MJF) printing. The layer-by-layer fabrication process of MJF not only improves the mechanical and flame-retardant properties of the printed GC sensor but also bolsters its robustness and sensitivity. The direction of sensor bending significantly impacts the relative resistance changes, allowing for precise investigations of joint motions in the human body, such as those of the fingers, wrists, elbows, necks, and knees. Furthermore, the data of resistance changes collected by the GC sensor are utilized to train asupport vector machine with a 95.83% accuracy rate for predicting human motions. Due to its stable humidity sensitivity, the sensor also demonstrates excellent performance in monitoring human breath and predicting breath modes (normal, fast, and deep breath), thereby expanding its potential applications in healthcare. This work opens up new avenues for using MJF-printed wearable sensors for a variety of healthcare applications.

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.

Deep learning-based activity-aware 3D human motion trajectory prediction in construction

Predicting human motion is a critical requirement in various applications, with particular significance in the construction sector. This task presents significant challenges due to the diverse nature of human actions and the complexities of converting 2D image coordinates to real-world space. In response to these challenges, this paper introduces an innovative deep learning-driven approach to forecast human motion trajectories, with a novel emphasis on activity recognition to improve predictive accuracy. The method utilizes a multi-camera system to extract 2D joint locations, which are then fused using a particle filter technique for 3D pose generation. Using 3D data, deep learning models are developed to first recognize the activity class, and then take it as auxiliary information for predicting the motion trajectory. Through a comprehensive experiment, we evaluated the proposed methodology. While the main innovation of the proposed approach lies in the incorporation of deep learning-based activity recognition into the trajectory prediction system, the experiment results revealed the activity-aware system’s capability to enhance prediction performance by a minimum of 6.4% and up to 16.6% in short-term forecasts, in compare with a conventional approach. Additionally, we analyzed the effects of varying time windows and joint selections on predictive outcomes across diverse scenarios and discussed the implications of these findings. By enhancing the prediction of human motion, this approach holds promise in improving workspace safety while encouraging effective interactions within complex environments.

Alternate method to develop ground motion prediction equations: Calibration over Himalayan region

An alternative approach for constructing GMPEs, the Consistent Spectral Shape approach (CSS), is recently proposed to overcome the limitations of the conventional methods. The CSS framework is already implemented using the NGA-West2 database to develop a set of five GMPEs of 5 % damped spectral accelerations (RotD50, RotD100, Geo-mean, GMRotD50, and GMRotD100). In this paper, the performance of the CSS-GMPEs based on the NGA-West2 database is first assessed against those developed in the NGA-West2 project. Two different yardsticks are considered for the comparison: overall root mean square error (RMSE) and similitude of spectral shape. RMSE for CSS-GMPE and five NGA-West2 GMPEs are nearly consistent, but CSS utilizes a significantly smaller number of independent variables. CSS-GMPE also performs better while comparing the similitude of spectral shape. This paper next modifies the existing set of five CSS-GMPEs for the Himalayan region using a generic modification function contingent on intensity measure (RotD50, RotD100, Geo-mean, GMRotD50, and GMRotD100). All five CSS-GMPEs modified for the Himalayan region are observed to be performed satisfactorily for all practical purposes. These modified GMPEs are expected to be a valuable tool for site-specific seismic hazard analysis.

Wave-frequency and low-frequency motions of a deep-draft spar buoy in irregular waves based on a consistent second-order theory

This study investigates the effect of horizontal low-frequency (LF) displacements and velocities on the responses of floating structures in irregular waves, with a focus on a deep-draft spar buoy. It employs a consistent second-order hydrodynamic model in the time domain and its two variations to assess the influence of LF motions. To accurately estimate LF velocities from total velocities, an improved wavelet low-pass filter is proposed and applied, overcoming challenges posed by one-sided data from simulations. For the examined spar buoy, the incorporation of LF displacements and velocities in the seakeeping analysis is deemed essential, as they are found to contribute to the reduction of slowly varying, and consequently, total surge and pitch responses. The study finds that the standard deviations of LF surge and pitch motions scale with significant wave height Hs as (Hs)b, with b close to 4/3 for surge motion, highlighting viscous damping as the dominating damping mechanism. For the storm seastate, third-order viscous drag loads may also have made a non-negligible contribution. Meanwhile, pitch motion exhibits b values around 1.1 for both low and storm seastates, indicating the significance of both wave-drift and viscous damping in LF pitch motion prediction. The effects of mooring stiffness are also discussed.

Fault detection with confidence level evaluation for perception module of autonomous vehicles based on long short term memory and Gaussian Mixture Model

The reliability of perception is crucial for developing higher levels of autonomy to enhance the safety and performance of fully autonomous vehicles. As a result, fault detection in the perception module is an essential function for autonomous vehicles. While many approaches rely on hardware redundancy, necessitating additional sensors, achieving analytical redundancy in perception is a pivotal factor in reducing the cost and complexity of the autonomous system. This paper utilizes a motion predictor for surrounding vehicles and a Gaussian Mixture Model (GMM) to evaluate the confidence level of the perception module without requiring additional sensor installations. The motion predictor is designed based on a long short-term memory-based recurrent neural network. The error between the motion prediction and detection results is leveraged to estimate the confidence level of the perception module. In essence, the motion prediction results are treated as redundant sensor measurements. The error distribution is modeled using a GMM, and the cumulative probability density of the GMM is employed to assess the confidence level. The proposed algorithm's effectiveness is evaluated through a driving data-based simulation study with fault injection. This study shows an enhanced sensitivity to injected faults and a reduced time for fault detection. Furthermore, the proposed algorithm provides a quantitative estimate of the performance degradation level of the perception module. This estimation can serve as an indicator of uncertainty that the motion planner should account for.

An end-to-end neural network for UUV autonomous collision avoidance

This paper proposes a 3D autonomous collision avoidance method based on convolutional gated recurrent units to improve the autonomy of Unmanned Underwater Vehicle (UUV). The state equations of the UUV autonomous collision avoidance system are constructed by studying its mechanism and integrating dynamic/static obstacle recognition, dynamic obstacle motion prediction, collision risk assessment, and collision avoidance. Then a multi-input single-output neural network architecture that integrates static feature extraction, dynamic time sequence modeling, and feature integration is proposed based on Convolutional Neural Networks (CNNs) and Gated Recurrent Units (GRUs) to describe the state space. CNNs extract features from sonar observation data to improve the accuracy of obstacle recognition. GRUs are combined with CNNs to capture the correlation of long-distance features and extract dynamic features. The spatial and temporal invariance of the neural network architecture enhances the fault tolerance of the UUV collision avoidance system for inputs and adaptability to observation noise and environments. Finally, simulation results show that this method is adaptable to sonar observation noise and unknown environments to solve the problem of forward-looking sonar-based UUV collision avoidance in unknown complex ocean environments.

Ground motion prediction equation for NW Himalaya and its surrounding region

An attenuation relationship, also known as a Ground Motion Prediction Equation (GMPE), has been developed for the Northwest Himalaya and its surrounding region. This GMPE utilizes strong motion data recorded by two networks: the National Center for Seismology Network and the strong motion network installed by Indian Institute of Technology Roorkee named PESMOS. For the development of this GMPE, 309 strong motion records from 86 earthquakes that occurred between 2005 and 2021 were considered. The selected earthquakes have magnitudes ranging from 3.0 to 6.9. In this study, the average of Peak Ground Acceleration (PGA) of two horizontal components was used to develop the new attenuation relationship. A multi-regression model was chosen for the analysis. The proposed attenuation relationship is.Log10 PGA = 1.889 + 0.3996 × M - 0.95736 log10 (HD + exp (0.4114 × M)) ± 0.3646Where, PGA is peak horizontal acceleration in gal of strong ground motion, M is the magnitude, and HD is the hypocentral distance from the source. The standard error of the proposed relationship is 0.3646. This newly developed attenuation relationship holds significance for various applications, including site-specific studies, seismic hazard estimation, ground motion simulation, earthquake early warning (EEW), and engineering applications. It provides valuable insights into the behavior of ground motion in the Northwest Himalayan region, allowing for better preparedness and mitigation strategies against seismic hazards in the area.

Uncertainties of 3-D Motion Prediction From a Single Virtual Accelerometer During Walking and Running

Uncertainties of 3-D Motion Prediction From a Single Virtual Accelerometer During Walking and Running

Flow-Based Spatio-Temporal Structured Prediction of Motion Dynamics.

Conditional Normalizing Flows (CNFs) are flexible generative models capable of representing complicated distributions with high dimensionality and large interdimensional correlations, making them appealing for structured output learning. Their effectiveness in modelling multivariates spatio-temporal structured data has yet to be completely investigated. We propose MotionFlow as a novel normalizing flows approach that autoregressively conditions the output distributions on the spatio-temporal input features. It combines deterministic and stochastic representations with CNFs to create a probabilistic neural generative approach that can model the variability seen in high-dimensional structured spatio-temporal data. We specifically propose to use conditional priors to factorize the latent space for the time dependent modeling. We also exploit the use of masked convolutions as autoregressive conditionals in CNFs. As a result, our method is able to define arbitrarily expressive output probability distributions under temporal dynamics in multivariate prediction tasks. We apply our method to different tasks, including trajectory prediction, motion prediction, time series forecasting, and binary segmentation, and demonstrate that our model is able to leverage normalizing flows to learn complicated time dependent conditional distributions.

Scenario-based model predictive control with probabilistic human predictions for human–robot coexistence

This paper proposes a real-time motion planning scheme for safe human–robot workspace sharing relying on scenario-based nonlinear model predictive control (NMPC), a well-known approach for solving stochastic NMPC problems. A scenario tree is generated via higher-order Markov chains to provide probabilistic predictions of the human motion. Scenario-based NMPC is then used to generate point-to-point motions of the robot manipulator based on the above-mentioned human motion predictions, accounting for safety constraints via speed and separation monitoring (SSM). This means that the robot speed is always modulated to be able to stop before a possible collision with the human occurs. After proving theoretical properties on recursive feasibility and closed-loop stability of the proposed motion planning strategy, this is tested experimentally on a Kinova Gen3 robot interacting with a human operator, showing superior performance with respect to an NMPC scheme not relying on human predictions and to a fixed-path SSM strategy.

The effect of eccentricity on visual motion prediction in peripheral vision.

The purpose of the current study was to clarify the effect of eccentricity on visual motion prediction using a time-to-contact (TTC) task. TTC indicates the predictive ability to accurately estimate the time-to-contact of a moving object based on visual motion perception. We also measured motion reaction time (motion RT) as an indicator of the speed of visual motion perception. The TTC task was to press a button when the moving target would arrive at the stationary goal. In the occluded condition, the target dot was occluded 500 ms before the time to contact. The motion RT task was to press a button as soon as the target moved. The visual targets were randomly presented at five different eccentricities (4°, 6°, 8°, 10°, 12°) and moved on a circular trajectory at a constant tangent velocity (8°/s) to keep the eccentricity constant. Our results showed that TTC in the occluded condition showed an earlier response as the eccentricity increased. Furthermore, the motion RT became longer as the eccentricity increased. Therefore, it is most likely that a slower speed perception in peripheral vision delays the perceived speed of motion onset and leads to an earlier response in the TTC task.

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.

Model selection for prediction of strong ground motion peaks in Türkiye

Model selection for prediction of strong ground motion peaks in Türkiye

Ambient noise differential adjoint tomography reveals fluid-bearing rocks near active faults in Los Angeles

Water scarcity is a pressing issue in California. We develop ambient noise differential adjoint tomography that improves the sensitivity to fluid-bearing rocks by canceling bias caused by noise sources. Here we image the shallow S-wave velocity structure using this method beneath a linear seismic array (LASSIE) in Los Angeles Basin, which shows significant velocity reduction marking a major regional water producer, the Silverado aquifer, along with other fluid-bearing structures. Based on the S-wave tomography and previous P-wave studies, we derive the porosity in Long Beach and discover that the rock from 1-2 km depth surrounding the Newport-Inglewood Fault contains abundant fluids with pore-fluid fraction ~0.33. The high-porosity rock around the fault coincides with previously observed week-long shallow seismicity south of LASSIE array in Long Beach. The imaged S-wave velocity in the top layer shows a similar trend in the geotechnical layer Vs 30, suggesting additional applications to ground motion prediction.

Development of Synthetic Ground Motion-based Attenuation Relationship for Bihar Region for Seismic Ground Response Analysis Considering Central Seismic Gap

Ground Motion Prediction Equation (GMPE) is one of the significantly important tools to perform the seismic hazards analysis of any region. Therefore, the development of GMPEs at the bedrock level is utmost important especially when the region does not have any earthquake recording stations. The present study discussed the development of a GMPE at bedrock level for the Bihar region based on the stochastic model. The different seismic parameters such as magnitudes (Mw) 4.0‐8.5, spectral periods of 0‐10 s and distances up to 300 km have been considered for the stochastic model. Based on the results, it was found that the stochastic model is capable to predict the ground motion synthetically and the proposed GMPE, for Bihar region, predicts the spectral acceleration in most precise way. Further, the ground motion amplification analysis was carried out using synthetically generated bedrock motion to analyze the effect of soil deposits on the amplification or de‐amplification of the bedrock peak ground acceleration. It was found that that the seismic wave gets amplified at ground level by 10% to 70% from the input motion PGA ranging from 0.175g‐0.435g, indicating amplification and de‐amplification of seismic wave. The maximum spectral acceleration at surface level was also found to be increased by approximately 60%, 56% and 27%, when bedrock input motion of PGA = 0.175g, 0.256g and 0.435g, respectively. Thus, based on the results, it can be stated that the developed GMPE can be used to assess the seismic hazards analysis in Bihar region. Further, it can be suggested that there is a need of the development of a predictive attenuation relationship at the surface level PGA, for Bihar region or any earthquake prone area, incorporating different site classes and regional seismicity since, the seismic wave amplified due to the presence of soil deposits.

Recent Developments and Future Directions of Wearable Skin Biosignal Sensors

AbstractThis review article explores the transformative advancements in wearable biosignal sensors powered by machine learning, focusing on four notable biosignals: electrocardiogram (ECG), electromyogram (EMG), electroencephalogram (EEG), and photoplethysmogram (PPG). The integration of machine learning with these biosignals has led to remarkable breakthroughs in various medical monitoring and human–machine interface applications. For ECG, machine learning enables automated heartbeat classification and accurate disease detection, improving cardiac healthcare with early diagnosis and personalized interventions. EMG technology, combined with machine learning, facilitates real‐time prediction and classification of human motions, revolutionizing applications in sports medicine, rehabilitation, prosthetics, and virtual reality interfaces. EEG analysis powered by machine learning goes beyond traditional clinical applications, enabling brain activity understanding in psychology, neurology, and human–computer interaction, and holds promise in brain–computer interfaces. PPG, augmented with machine learning, has shown exceptional progress in diagnosing and monitoring cardiovascular and respiratory disorders, offering non‐invasive and accurate healthcare solutions. These integrated technologies, powered by machine learning, open new avenues for medical monitoring and human–machine interaction, shaping the future of healthcare.

Application of neural network in object motion tracking in the context of big data

In the context of big data, the advancements in computer analytics and computational capability have promoted transformative opportunities. At the forefront of this progress, neural networks have emerged as an essential instrument across diverse domains, catalyzing enhancements in operational efficiency. This paper embarks on an exploration of the applications of neural networks within the realm of object motion prediction and tracking, with the primary objective of improving the efficacy and precision of tracking mechanisms. Amidst the escalating complexities of data proliferation, the imperative for streamlined tracking methodologies has intensified. In response, this study investigates the amalgamation of neural networks with object motion prediction and tracking processes. By using the powerful abilities of neural networks, the research aims to bring a significant change to the field. This change will be seen in improved accuracy and faster tracking efficiency. Combining neural networks with motion prediction and tracking creates a cooperative partnership that develops a new system based on data-driven insights. By merging computational power and advanced algorithms, the paper anticipates a fresh direction for tracking technology.

The Generation of the Target Aftershock Spectrum Based on the Conditional Mean Spectrum of Aftershocks

Numerous studies have examined the responses of various structures to the mainshock–aftershock (MS–AS) ground motion, and the MS–AS ground motions are very important as the input. Therefore, in the absence of aftershock information, it is particularly critical to construct a reasonable MS–AS seismic sequence. This paper aims to provide a new reasonable method for generating the target aftershock response spectrum, which can be used to select or artificially simulate aftershock ground motion, given the seismic information of the main shock. Firstly, the magnitude, fault size, and location of the aftershock are determined. Then, other parameters required for the aftershock ground motion prediction equation (GMPE) are calculated. Subsequently, the correlation of the spectral shape to the MS–AS ground motion is used to modify the response spectrum predicted using the GMPE to obtain the conditional mean spectrum of aftershocks (CMSA). Finally, the relative errors of the predicted spectrum via the ASK14 model and CMSA are compared for four different assumptions. The results show that the simulated aftershock parameters and the actual ones accord well, and the relative errors of the CMSA can be controlled within 20%. Meanwhile, the discrete property of the target aftershock response spectrum is closer to the real recorded response spectrum.