Motion Generation Research Articles

A learnable motion preserving pooling for action recognition

Using deep neural networks (DNN) for video understanding tasks is expensive in terms of computation cost. Pooling layers in DNN which are widely used in most vision tasks to resize the spatial dimensions play crucial roles in reducing the computation and memory cost. In video-related tasks, pooling layers are also applied, mostly in the spatial dimension only as the standard average pooling in the temporal domain can significantly reduce its performance. This is because conventional temporal pooling degrades the underlying important motion features in consecutive frames. Such a phenomenon is rarely investigated and most state-of-art methods simply do not adopt temporal pooling, leading to enormous computation costs. In this work, we propose a learnable motion-preserving pooling (MPPool) layer that is able to preserve the general motion progression after the pooling. This pooling layer first locates the frames with the strongest motion features and then keeps these crucial features during pooling. Our experiments demonstrate that MPPool not only reduces the computation cost for video data modeling, but also increases the final prediction accuracy on various motion-centric and appearance-centric datasets.

Neuromuscular synergy characteristics of Tai Chi leg stirrup movements: optimal coordination patterns throughout various phases

ObjectiveTo investigate the neuromuscular activity characteristics of Tai Chi athletes and identify optimal muscle synergy patterns.MethodData were collected from 12 elite Tai Chi athletes using a Vicon motion capture system, a Kistler 3D force plate, and a Noraxon surface electromyography system. Muscle synergy patterns were extracted using Non-negative Matrix Factorization.ResultsFour muscle synergy patterns were identified in each of the three phases of the leg stirrup movement, with the optimal synergy pattern for each phase determined as follows: knee lift phase: rectus femoris and vastus lateralis of the right leg; extension phase: rectus femoris, vastus lateralis, biceps femoris, and medial gastrocnemius of the right leg; recovery phase: rectus femoris, vastus lateralis, and medial gastrocnemius of the right leg. These patterns explain the muscle coordination activities for each phase.ConclusionThis study identified the optimal muscle synergy patterns for each phase, supporting the fluidity and force generation of the leg stirrup movement. This provides Tai Chi athletes with a more efficient way to exert strength and maintain balance.

The Psychological Nature of Female Gait Attractiveness.

Walking, a basic physical movement of the human body, is a resource for observers in forming interpersonal impressions. We have previously investigated the expression and perception of the attractiveness of female gaits. In this paper, drawing on our previous research, additional analysis, and reviewing previous studies, we seek to deepen our understanding of the function of gait attractiveness. First, we review previous research on gait as nonverbal information. Then, we show that fashion models' gaits reflect sociocultural genderlessness, while nonmodels express reproductive-related biological attractiveness. Next, we discuss the functions of gait attractiveness based on statistical models that link gait parameters and attractiveness scores. Finally, we focus on observers' perception of attractiveness, constructing a model of the visual information processing with respect to gait attractiveness. Overall, our results suggest that there are not only biological but also sociocultural criteria for gait attractiveness, and men and women place greater importance on the former and latter criteria, respectively, when assessing female gait attractiveness. This paper forms a major step forward in neuroaesthetics to understand the beauty of the human body and the generation of biological motions.

Synthesis and analysis of Stephenson III six-bar motion generators with an instantaneous stop link

Synthesis and analysis of Stephenson III six-bar motion generators with an instantaneous stop link

Analysis of a Dry Friction Force Law for the Covariant Optimal Control of Mechanical Systems with Revolute Joints

This contribution shows a geometric optimal control procedure to solve the trajectory generation problem for the navigation (generic motion) of mechanical systems with revolute joints. The mechanical system is analyzed as a nonlinear Lagrangian system affected by dry friction at the joint level. Rayleigh’s dissipation function is used to model this dissipative effect of joint-level friction, and regarded as a potential. Rayleigh’s potential is an invariant scalar quantity from which friction forces derive and are represented by a smooth model that approaches the traditional Coulomb’s law in our proposal. For the optimal control procedure, an invariant cost function is formed with the motion equations and a Riemannian metric. The goal is to minimize the consumed energy per unit time of the system. Covariant control equations are obtained by applying Pontryagin’s principle, and time-integrated using a Finite Elements Method-based solver. The obtained solution is an optimal trajectory that is then applied to a mechanical system using a proportional–derivative plus feedforward controller to guarantee the trajectory tracking control problem. Simulations and experiments confirm that including joint-level friction forces at the modeling stage of the optimal control procedure increases performance, compared with scenarios where the friction is not taken into account, or when friction compensation is performed at the feedback level during motion control.

The Synthesis of Spherical Four-bars for Biomimetic Motion through Complete Solutions for Approximate Rigid Body Guidance

Abstract In this paper, we form a constrained optimization problem for spherical four-bar motion generation. Instead of using local optimization methods, all critical points are found using homotopy continuation solvers. The complete solution set provides a full view of the optimization landscape and gives the designer more freedom in selecting a mechanism. The motion generation problem admits 61 critical points, of which two must be selected for each four-bar mechanism. We sort solutions by objective value and perform a second order analysis to determine if the solution is a minimum, maximum, or saddle point. We apply our approximate synthesis technique to two applications: a hummingbird wing mechanism and a sea turtle flipper gait. Suitable mechanisms were selected from the respective solution sets and used to build physical prototypes.

Hybrid Orientation and Position Collaborative Motion Generation Scheme for a Multiple Mobile Redundant Manipulator System Synthesized by a Recurrent Neural Network.

To enable distributed multiple mobile manipulator systems to complete collaborative tasks safely and stably, this article investigates and presents a motion generation scheme that considers both orientation and position coordination based on a distributed recurrent neural network. Moreover, physical limits are also considered. Specifically, the orientation and position coordination constraints and physical limits are modeled separately as equality and inequality constraints with coupled variables. Subsequently, a motion generation scheme for multiple mobile manipulators based on quadratic programming is established. Finally, a distributed linear variational inequality-based primal-dual neural network is constructed to solve the motion generation scheme and obtain the motion trajectories of all the mobile manipulators. The simulation results demonstrate that the hybrid orientation and position collaboration motion generation scheme effectively addresses the position and orientation coordination problem for multiple mobile manipulator systems. Compared to other schemes, the proposed scheme based on a distributed computing structure greatly enhances the stability of the system. Additionally, the proposed approach introduces orientation coordination and physical limits, which increases the practicality of the system.

Learning Constrained Dynamic Correlations in Spatiotemporal Graphs for Motion Prediction.

Human motion prediction is challenging due to the complex spatiotemporal feature modeling. Among all methods, graph convolution networks (GCNs) are extensively utilized because of their superiority in explicit connection modeling. Within a GCN, the graph correlation adjacency matrix drives feature aggregation, and thus, is the key to extracting predictive motion features. State-of-the-art methods decompose the spatiotemporal correlation into spatial correlations for each frame and temporal correlations for each joint. Directly parameterizing these correlations introduces redundant parameters to represent common relations shared by all frames and all joints. Besides, the spatiotemporal graph adjacency matrix is the same for different motion samples, and thus, cannot reflect samplewise correspondence variances. To overcome these two bottlenecks, we propose dynamic spatiotemporal decompose GC (DSTD-GC), which only takes 28.6% parameters of the state-of-the-art GC. The key of DSTD-GC is constrained dynamic correlation modeling, which explicitly parameterizes the common static constraints as a spatial/temporal vanilla adjacency matrix shared by all frames/joints and dynamically extracts correspondence variances for each frame/joint with an adjustment modeling function. For each sample, the common constrained adjacency matrices are fixed to represent generic motion patterns, while the extracted variances complete the matrices with specific pattern adjustments. Meanwhile, we mathematically reformulate GCs on spatiotemporal graphs into a unified form and find that DSTD-GC relaxes certain constraints of other GC, which contributes to a better representation capability. Moreover, by combining DSTD-GC with prior knowledge like body connection and temporal context, we propose a powerful spatiotemporal GCN called DSTD-GCN. On the Human3.6M, Carnegie Mellon University (CMU) Mocap, and 3D Poses in the Wild (3DPW) datasets, DSTD-GCN outperforms state-of-the-art methods by 3.9%-8.7% in prediction accuracy with 55.0%-96.9% fewer parameters. Codes are available at https://github.com/Jaakk0F/DSTD-GCN.

Conditional Adversarial Motion Priors by a Novel Retargeting Method for Versatile Humanoid Robot Control

ABSTRACTSignificant advancements have been made in the field of humanoid robot, particularly in walking control strategies. However, achieving straight‐legged walking remains a challenge. Both the traditional model‐based and the learning‐based control methods confront with difficulties in achieving natural humanoid gait feature. To address this issue, a general motion retargeting method is developed and also evaluated for humanoid robots with different structure, size and degrees of freedom. Moreover, a conditional adversarial motion priors method is proposed based on reinforcement learning and validated on the humanoid robot GTX‐III. Through various motion segments from the motion capture database, it is shown that this method can successfully enable the humanoid robot to perform straight‐legged walking with flexible and natural transitions between different gaits within a single discriminator network.

Brownian motion in a vector space over a local field is a scaling limit

For any natural number d, the Vladimirov–Taibleson operator is a natural analogue of the Laplace operator for complex-valued functions on a d-dimensional vector space V over a local field K. Just as the Laplace operator on L2(Rd) is the infinitesimal generator of Brownian motion with state space Rd, the Vladimirov–Taibleson operator on L2(V) is the infinitesimal generator of real-time Brownian motion with state space V. This study deepens the formal analogy between the two types of diffusion processes by demonstrating that both are scaling limits of discrete-time random walks on a discrete group. It generalizes the earlier works, which restricted V to be the p-adic numbers.

Electronically actuated artificial hinged cilia for efficient bidirectional pumping.

Cilial pumping is a potent mechanism used to control and manipulate fluids on microscales. Recently, we introduced an electronically driven μ-cilial platform that can create arbitrary flow patterns in liquids near a surface with the potential for various engineering applications. This μ-cilial platform, however, utilized the coupling between elasticity and viscous drag to obtain pumping and had several limitations. For example, each cilium could only pump in one direction. Thus, to create bidirectional flows, it was necessary to fabricate and separately actuate two oppositely facing cilia. As another example, the generation of non-reciprocal cilial motions, a necessary condition for pumping at these scales, could only be achieved by matching the elastic stresses inherent in actuating the cilia with the viscous drag forces generated by the flows. This criterion severely restricted the frequency range over which the cilia could be operated and resulted in a small swept area, both of which restricted the volume of fluid being pumped in each cycle. These limitations contrast with the capabilities of natural cilia, which can achieve omnidirectional transport and operation over a broad range of frequencies. In natural cilia, these capabilities arise from their complex internal structure. Inspired by this strategy we designed hinged cilia and show they can achieve bidirectional pumping of larger fluid volumes over a broad range of frequencies. Finally, we demonstrate that even regular arrays of individually controlled hinged cilia can generate a variety of flow patterns using fewer cilia than in previous cilia metasurface designs.

Sensor‐Embedded Muscle for Closed‐Loop Controllable Actuation in Proprioceptive Biohybrid Robots

Biohybrid robots are soft robots that exploit unique characteristics of biological cells and tissues for motion generation. Skeletal muscle tissue‐based bioactuators respond to externally applied stimuli, such as electrical fields. However, current bioactuation systems rely on open‐loop control strategies that lack knowledge of the actuator's state. The regulation of output force and position of biohybrid robots requires self‐sensing control systems that combine bioactuators with sensors and control paradigms. Herein, a soft, fiber‐shaped mechanical sensor based on a piezoresistive composite is proposed that efficiently integrates with engineered skeletal muscle tissue and senses its contracting states in a cell culture environment in the presence of applied electrical fields. After testing the sensor's insulation and biocompatibility, its sensitivity for typical strains (<1%) is characterized, and its ability is proven to detect motions from contractile skeletal muscle tissue constructs. Finally, it is shown that the sensor response can feed an autonomous control system, thus demonstrating the first proprioceptive biohybrid robot that senses and responds to its contraction state. In addition to inspiring implantable systems, biomedical models, and other bioelectronic devices, the proposed technology will confer biohybrids with decisional autonomy, thus driving the paradigm shift between bioactuators and intelligent biohybrid robots.

Gaze modeling in multi-party dialogues and extraversion expression through gaze aversion control

Gaze control plays a crucial role in the context of conversational robots involving multiple participants. Particularly, gaze aversion is not only important for improving human likeness but is also strongly associated with the expression of a robot's personality and should be accounted for dialogue interactions. While previous studies on gaze models considering multi-party conversations have been conducted, those considering gaze aversion are limited. In this study, we developed a gaze motion generation model, focusing on three features: (1) the gaze target (towards dialogue partners or gaze aversion), (2) the gaze duration, and (3) the eyeball direction during gaze aversion. We extracted the gaze model parameters of two individuals with distinct extraversion from a multimodal three-party dialogue database and created a model that can modulate extraversion expression by setting a gaze aversion ratio parameter. We implemented this model on an android robot, which can reproduce human-like gaze behaviors. Video-based experiments have been conducted where participants watched videos of the robot talking with two humans, replicating the three-party dialogue scenarios used in this study, and evaluated the impressions on the robot's extraversion. The results demonstrated that by controlling gaze aversion parameters, the robot could exhibit various levels of extraversion.

Design and Modeling of a Long Range Motion Compliant Nanopositioning Stage Driven by a Normal Stressed Electromagnetic Actuator

Compliant nanopositioning stages with built-in ultra-precision actuators are frequently integrated into production and analysis instruments comprising ultra-high precision motion generation systems. These stages are essential nanotechnology and advanced material analysis components, providing precise positioning capabilities for various applications. However, in the practical engineering field, there is a lack of compliant nanopositioning stages that can achieve a long-range motion while maintaining accuracy, reliability, and compact size, which is the inspiration for this research. This paper investigates the design, modeling, and experimental testing of a long-range motion-compliant nanopositioning stage driven by a normal stressed electromagnetic actuator (NSEA). The nanopositioning stage components’ structural framework and working principle, including NSEA, bridge type distributed compliant (BTDC) mechanism, and the guiding mechanism, are fully examined to derive an analytical model. The analytical model is utilized in the sections that follow. Factors affecting the stroke and natural frequency of the nanopositioning stage are also illustrated. The optimization process of the nanopositioning stage is conducted in pursuit of a high-precision stage by specifically looking into the electromagnetic, BTDC mechanism, and guiding mechanism parameters. This optimization procedure also takes into account various design constraints, including stiffness, saturation flux density, and stress. Furthermore, the finite element analysis is used to verify the analytical model, and the results are discussed. The prototype is fabricated with reference to the analytical and finite element analysis results, and the experimental tests are conducted, including motion and natural frequency tests. In addition, a control system, which adopts both a proportional-integral-derivative controller and a damping controller, is designed to create a closed-loop system. Finally, the tracking performance of the stage was investigated, and a very minimal tracking error was observed. Overall, the comprehensive models and experimental tests proved the stage to be a good model which achieved the objective of the research.

Entropy generated nonlinear mixed convective beyond constant characteristics nanomaterial wedge flow

Background and objectiveNanomaterial flows at present have received much attention of the researchers. Such motivation stems for their utilization in pharmaceutical, industrial, petroleum, manufacturing and engineering applications including regenerative medicine, treatment of cancer, electronic device cooling, solar collector, mineral oil, antimicrobial agents, thermal storage system, transformer cooling and X-imaging etc. Higher thermal energy requirement in several electrical and mechanical systems is desired in recent time due to high advancement of science and technology. For important application here the stagnation point flow due to wedge with nonlinear convection is explored. For porous media, the Darcy-Forchheimer relation is considered. Characteristics of nanofluid is added by utilizing Buongiorno's model. Variable fluid characteristics are under consideration. Applied magnetic field is accounted. Energy expression comprises thermal radiation, Brownian motion, Ohmic heating, thermophoresis and heat generation. First order reaction and entropy rate are considered. MethodologyNonlinear differential expressions are changed into dimensionless ordinary systems through adequate transformations. The resultant nonlinear ordinary systems are solved through optimal homotopy analysis technique (OHAM). ResultsOutcomes for influential variables about temperature, velocity, concentration and entropy rate have been arranged. Here results show that for magnetic field the liquid flow and entropy rate have opposite scenarios. Thermal distribution and concentration have reverse effects for random motion variable while similar effect holds for thermophoresis parameter. Magnetic field and diffusion parameters contribute to augment entropy generation. Higher values of temperature dependent electrical conductivity variable lead to increase entropy rate. Thermal distribution for radiation and Eckert number is similar.

Three-Dimensional Lorentz-Invariant Velocities

Lorentz invariance underlies special relativity, and the energy formula and relative velocity formula are well known to be invariant under a Lorentz transformation. Here, we determine the functional forms in terms of four arbitrary functions for those three dimensional velocity fields that are automatically invariant under the most general fully three-dimensional Lorentz transformation. For general three-dimensional motion, using rectangular Cartesian coordinates (x,y,z), we determine the first-order partial differential equations for the three velocity components u(x,y,z,t), v(x,y,z,t) and w(x,y,z,t) in the x−, y− and z−directions respectively. These partial differential equations and the associated partial differential relations connecting energy and momentum are fully compatible with the Lorentz-invariant energy–momentum relations and appear not to have been given previously in the literature. We determine the spatial and temporal dependence of the functional forms for those three-dimensional velocity fields that are automatically invariant under three-dimensional Lorentz transformations. An interesting special case gives rise to families of particle paths for which the magnitude of the velocity is the speed of light. This is indicative of the abundant possibilities existing in the “fast lane”.

Hybrid Cooperative Control of Functional Electrical Stimulation and Robot Assistance for Upper Extremity Rehabilitation.

Hybrid systems that integrate Functional Electrical Stimulation (FES) and robotic assistance have been proposed in neurorehabilitation to enhance therapeutic benefits. This study focuses on designing a cooperative controller capable of distributing the required torque for movement between robotic actuation and FES, thereby eliminating the need for time-consuming calibration procedures. The control schema comprises three main blocks: a motion generation block that defines the desired trajectory, a motor control block including both a weight compensation feedforward and a feedback impedance controller, and an FES control block, based on trial-by-trial Iterative Learning Control (ILC), that adjusts the stimulation intensity according to a predefined stimulation waveform. The feedforward motor assistance can be dynamically regulated using an allocation factor. Experiments involving 12 healthy volunteers were conducted using a one-degree-of-freedom elbow testbed. The experimental results showcased the successful integration of Functional Electrical Stimulation (FES) with robotic actuation, ensuring precise trajectory tracking with a Root Mean Square Error (RMSE) below 7°. Notably, allocating more torque to FES led to a 51 % reduction in motor torque. In conditions where FES operated alone, there was poorer tracking performance with an RMSE of 24° and an early onset of muscle fatigue, as evidenced by a reduced number of achieved repetitions. Furthermore, the hybrid approach enabled 100 fatigue-free elbow flexion repetitions, underscoring the effectiveness of cooperative FES-motor control in extending the benefits of FES-induced exercises. This study proposes a flexible approach which can be extended to a multi-degree-of-freedom hybrid system. Furthermore, it underscores the significance of employing a straightforward and adaptable methodology with a rapid calibration procedure, making it easily transferable to clinical applications.

An Empirical Spatiotemporal Input Model for Regional Seismic Ground Motions

The development of the economy and technology has attracted greater attention to regional seismic resistance, a key technology for which is the simulation of regional seismic ground motions. This study proposes a simple and highly operable method for the generation of regional seismic ground motions, which includes five steps. First, a seismic ground motion recorded at a seismic station is selected as the original ground motion. The original ground motion is then decomposed into eight sub-signals by wavelet packet transform (WPT). Next, the eight sub-signals are adjusted by the frequency attenuation model. The generated ground motion is then amplitude modulated by the peak ground acceleration (PGA) attenuation model. The frequency attenuation model and the PGA attenuation model are obtained by fitting seismic data from the 1999 Chi-Chi earthquake. Finally, the propagation of the ground motion is considered. The proposed method can provide a time history of seismic ground motions for any point in a region. To investigate the rationality and applicability of the proposed method, the characteristics of the generated seismic ground motions are compared with those of true seismic ground motions. The average error of the PGA of the generated ground motions and the true ground motions is 0.072[Formula: see text][Formula: see text]; thus, the proposed method is suitable for use in small-scale regions. Furthermore, 4-story and 15-story structures are used for structural response analysis. The story drift ratio error between the generated ground and true ground motions is found to be within an acceptable error range. The proposed method provides a new way to generate ground motions for regional seismic investigation.

A stochastic approach for the generation and combination of vertical and horizontal strong ground motion

Damage to structures during near-fault earthquakes often indicates that the vertical component was larger than what is usually considered in seismic design. Commonly, the vertical component of seismic ground motions is associated with the P-wave. In contrast, the horizontal seismic ground motions’ component is usually related to the arrival of the S-wave. However, findings of recent near-fault events indicate that both the horizontal and the vertical motions are a mixture of S and P-waves. Therefore, the need to generate three-component artificial accelerograms for structural design arises. This paper proposes a double aim: starting from the well-noted Snell’s law, a physical explanation of the mixture of S and P-waves is outlined. Then, a novel integrodifferential filter system for the stochastic simulation of the three components of strong motion is presented. The proposed study is applied to some theoretical applications as well as to two real Italian seismic events.

Carreau nanofluid dynamics with activation energy gyrotactic microorganisms in a porous medium: Application to solar energy

This study investigates the magnetohydrodynamic (MHD) flow of Carreau nanofluid through a porous medium with motile microorganisms, focusing on various geometries under shear-thinning and shear-thickening conditions. The aim is to elucidate how factors such as activation energy, Schmidt number, Peclet number, bioconvection, Brownian motion, thermophoresis, and heat generation influence flow dynamics. Using similarity transformations, we nondimensionalize the governing equations and solve them numerically with the Runge–Kutta method and a shooting technique in MATLAB. Our findings indicate that variations in Carreau, magnetic, and suction parameters notably impact velocity, temperature, concentration, diffusion, wall friction, and heat transfer, generally resulting in reduced values. Specifically, the flat plate geometry exhibits lower skin friction, heat transfer, and mass transfer rates, as well as decreased gyrotactic microorganism effects. Increased activation energy enhances concentration fields, signaling slower chemical reactions, while higher Peclet numbers and bioconvection inversely affect flow properties. Additionally, reduced Schmidt numbers lead to lower microorganism concentrations. These results provide valuable insights into the complex interactions between fluid dynamics and microorganism behavior, with implications for optimizing processes in biotechnology and environmental management.