Force Information Research Articles

ænet-PyTorch: A GPU-supported implementation for machine learning atomic potentials training.

In this work, we present ænet-PyTorch, a PyTorch-based implementation for training artificial neural network-based machine learning interatomic potentials. Developed as an extension of the atomic energy network (ænet), ænet-PyTorch provides access to all the tools included in ænet for the application and usage of the potentials. The package has been designed as an alternative to the internal training capabilities of ænet, leveraging the power of graphic processing units to facilitate direct training on forces in addition to energies. This leads to a substantial reduction of the training time by one to two orders of magnitude compared to the central processing unit implementation, enabling direct training on forces for systems beyond small molecules. Here, we demonstrate the main features of ænet-PyTorch and show its performance on open databases. Our results show that training on all the force information within a dataset is not necessary, and including between 10% and 20% of the force information is sufficient to achieve optimally accurate interatomic potentials with the least computational resources.

Experimental research and numerical simulation based on granary effect

Based on Janssen model and discrete element model, experiment and numerical simulation are used to verify the granary effect of coal storage silo in power plant. In the experiment, the bottom pressure of two different particles in a variable-diameter cylinder is measured. The type of particles and the diameter of the cylinder are used as the research objects to explore the function relationship between the bottom pressure of the cylinder and the particle filling quality. In the experiment with soybeans as filling particles, the granary effect is more obvious, and the smaller the cylinder diameter is, the more obvious the granary effect is. At the same time, establish a mathematical model of particles, simulation of particle accumulation process, solve the force information of particles. Through numerical simulation, it is found that the data is more in line with the Janssen model, and the granary effect is more obvious.

Construction of a novel digital method for quantitative analysis of occlusal contact and force

BackgroundOcclusal analysis is essential in the dental clinical practice. However, the traditional occlusal analysis performed on the two-dimensional level can not directly correspond to the tooth surface with three-dimensional profile, therefore the clinical guidance value is limited.MethodsBy combining the 3D digital dental models and quantitative data from 2D occlusal contact analysis, this study constructed a novel digital occlusal analysis method. The validity and reliability of DP and SA were verified by comparing the results of occlusal analysis of 22 participants. ICC values for occlusal contact area (OCA) and occlusal contact number (OCN) were tested.ResultsResults confirmed the reliability of the two occlusal analysis methods with ICC values of 0.909 for SAOCA, 0.906 for DPOCA, 0.929 for SAOCN and 0.904 for DPOCN. The Bland-Altman plot, paired t-test (tOCN = 0.691, P > 0.05) and Pearson correlation analysis results (R = 0.68, p < 0.001) verified the validity between SA and DP. Then a novel digital occlusal analysis method was constructed, which not only can locate the occlusion contact and provide the quantitative analysis, but also provide a comprehensive description of the resultant force of each tooth and the component forces on the x-, y- and z-axis.ConclusionsThis new occlusal analysis method can obtain quantitative analysis of occlusal contact including contact area and force information simultaneously, which will provide new impetus and greater help for clinical dental treatment and scientific research.

A Miniature Multiaxis Force/Torque Sensor for Acupuncture

To realize the quantitative analysis and research of traditional Chinese medicine (TCM) acupuncture manipulations, miniaturized multidimensional force sensors are necessary to measure the lifting–thrusting force and twisting torque in the process of acupuncture, which records the manipulations of excellent physicians for relevant physicians to learn and improve their manipulations. A miniature multiaxis force/torque ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${F}/{T}$ </tex-math></inline-formula> ) sensor for acupuncture is developed in this article. The elastomer of the design adopts hollow thin-wall cross beam structure with inverted <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${T}$ </tex-math></inline-formula> floating beam, which retains the space for the strain gauge to be pasted, reduces the coupling between dimensions, and improves the sensitivity and accuracy. This design has the advantages of miniaturization, lightweight, high precision, and convenient needle substitution. The elastomer is processed with polyetheretherketone (PEEK) material to reduce the mass. First, the feasibility of the hollow thin-wall cross beam structure with inverted <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${T}$ </tex-math></inline-formula> floating beam is verified by ANSYS finite element simulation. Then, the hardware system based on STM32 is designed to realize signal acquisition, amplification, and transmission. Next, the static calibration and decoupling of the sensor are accomplished with the calibration device. The calibration results indicate that the accuracy of the sensor is high (the coupling error is less than 2%). Finally, the acupuncture force measurement system is constructed to complete the acupuncture manipulation experiment. The experimental results demonstrate that our system can accurately obtain the needle force information and can distinguish different acupuncture manipulations, which verifies the feasibility of the sensor while realizing the quantitative characterization of acupuncture manipulations.

BUILDING COMPOSITES AS OBJECTS OF SYSTEM ANALYSIS

The article is devoted to the consideration of the material science problems from the standpoint of the general systems theory. The method of dialectical pairs "part-whole", "continuous-discrete" and "randomness-necessity" is used, corresponding to the general system coordinates, on the basis of which the third emergent component of a synthetic nature arises. System-theoretic methods make it possible to concretize model representations using elements of the crystallization theory of dynamic information theory. System representations are used to refine the concept of an interface in materials and heterogeneous systems. The dialectical unity of atomistic and geometric concepts in the theory of the formation of a new phase in binders is shown. The probabilistic and informational description is considered as a realization of the dialectical pair "randomness-necessity". The concept of information is closely related to choosing one or more options from many and remembering the choice made. The choice can be made as a result of the action of external forces (information reception) or due to the unstable behavior of the system (the emergence of new information). The process of structure formation can be considered as the dynamics of an information system. The dynamics of the information system should include interactions that provide any structure-forming particle with the possibility of transition between the areas of influence of stable states with subsequent attraction to one of them. The formation of structures in a binder from the position of dynamic information theory is considered as the reception of information. It is necessary to perform work, due to which the system will go into one of the stable states for the reception of information. The formation of new elements and structures ‒ pores, capillaries and cracks is considered as a transition of the system to its own potential minima ‒ the epigenetic landscape of the material as an emerging system. In this case, part of the energy will be lost in the process of dissipation, and the physical entropy will increase by a value exceeding the amount of information received. Such an effect can be considered as the "price" of the observed macroscopic process of self-organization. The involvement of systems theory also helps, on the basis of the principle of adequacy, to indicate the way for choosing a model object that is optimal for studying the processes of structural transformations of a material.

Position‐based force tracking adaptive impedance control strategy for robot grinding complex surfaces system

AbstractAs a key technology of robot grinding, force control has great influence on grinding effects. Based on the traditional impedance control, a position‐based force tracking adaptive impedance control strategy is proposed to improve the grinding quality of aeroengine complex curved parts, which considers the stiffness damping environmental interaction model, modifies the reference trajectory by a Lyapunov‐based approach to realize the adaptive grinding process. In addition, forgotten Kalman filter based on six‐dimensional force sensor is used to denoise the force information and a three‐step gravity compensation process including static base value calculation, dynamic zero update and contact force real‐time calculation is proposed to obtain the accurate contact force between tool and workpiece in this method. Then, to verify the effectiveness of the proposed method, a simulation experiment which including five different working conditions is conducted in MATLAB, and the experiment studying the deviation between the reference trajectory and the actual position is carried out on the robot grinding system. The results indicate that the position‐based force tracking adaptive impedance control strategy can quickly respond to the changes of environmental position, reduce the fluctuation range of contact force in time by modifying the reference trajectory, compensate for the defect of the steady‐state error of the traditional impedance control strategy and improve the surface consistency of machined parts.

Effects of Sensing Tactile Arrays, Shear Force, and Proprioception of Robot on Texture Recognition

In robotics, tactile perception is important for fine control using robot grippers and hands. To effectively incorporate tactile perception in robots, it is essential to understand how humans use mechanoreceptors and proprioceptors to perceive texture. Thus, our study aimed to investigate the impact of tactile sensor arrays, shear force, and the positional information of the robot's end effector on its ability to recognize texture. A deep learning network was employed to classify tactile data from 24 different textures that were explored by a robot. The input values of the deep learning network were modified based on variations in the number of channels of the tactile signal, the arrangement of the tactile sensor, the presence or absence of shear force, and the positional information of the robot. By comparing the accuracy of texture recognition, our analysis revealed that tactile sensor arrays more accurately recognized the texture compared to a single tactile sensor. The utilization of shear force and positional information of the robot resulted in an improved accuracy of texture recognition when using a single tactile sensor. Furthermore, an equal number of sensors placed in a vertical arrangement led to a more accurate distinction of textures during exploration when compared to sensors placed in a horizontal arrangement. The results of this study indicate that the implementation of a tactile sensor array should be prioritized over a single sensor for enhanced accuracy in tactile sensing, and the use of integrated data should be considered for single tactile sensing.

Robotic assembly strategy via reinforcement learning based on force and visual information

Since assembly tasks are frequently performed in a wide range of industries, there have been many efforts to develop robotic assembly strategies. However, robotic assembly is only applicable in structured environments wherein a target object is placed in a fixed position, because the occurrence of a large error substantially degrades performance. Thus, there is still a need for a generalized assembly strategy that can cope with a large position/orientation error regardless of the shape. To this end, this study presents an assembly strategy based on both the force and visual information. Specifically, the trajectory of the robot is obtained by combining the output of two neural-network-based trajectory generators that receive the force and image information, respectively, and then a deep reinforcement learning algorithm is applied to obtain the optimal strategy. In this process, imitation learning is applied to train the force-based network using the demonstration data collected with the suggested hand-guiding method, and the probability distribution of the feature is introduced in the image-based network to enable a robot to quickly adapt to assembly parts with different shapes. The performance of the proposed assembly strategy is experimentally verified using various peg-in-hole tasks, and the results confirm that the robot can successfully accomplish an assembly task regardless of the shapes of the assembly parts, even when the initial position/orientation error is large.

Automatic extraction method of force chain information and its application in the flow photoelastic experiment of granular matter

The force chain is the core of the multi-scale analysis of granular matter. Accurately extracting the force chain information among particles is of great significance to the study of particle mechanics and geological hazards caused by particle flow. However, in the photoelastic experiment, the precise identification of the branching points of force chains has not been effectively realized. Therefore, this study proposes an automatic extraction method of force chain key information. First, based on the Hough transform and the Euclidean distance, a particle geometric information identification model is established and geometric information such as particle circle center coordinates, radius, contact point location, and contact angle is extracted. Then, a particle contact force information identification model is established following the color gradient mean square method. The model realizes the rapid calibration and extraction of a large number of particle media contact force information. Next, combined with the force chain composition criterion and its quasilinear feature, an automatic extraction method of force chain information is established, which solves the problem of the accurate identification of the force chain branch points. Finally, in the photoelastic experiment of ore drawing from a single drawpoint, the automatic extraction method of force chain information is verified. The results show that the macroscopic distribution of force chains during ore drawing from a single drawpoint is left–right symmetrical. Strong force chains are mostly located on the two sides of the model but in small numbers and they mainly develop vertically. Additionally, the ends are mostly in a combination of Y and inverted Y shapes, while the middle is mostly quasilinear. Weak force chains are abundant and mostly distributed in the middle of the model, and develop in different directions. The proposed extraction method accurately extracts the force chain network from the photoelastic experiment images and dynamically characterizes the force chains of granular matter, which has significant advantages in particle geometry information extraction, force chain branch point discrimination, force chain retrieval, and force chain distribution and its azimuthal characterization. The results provide a scientific basis for studying the macroscopic and microscopic mechanical parameters of granular matter.

Effective haptic feedback type for robot-mediated material discrimination depending on target properties

Haptic feedback enables material perception via remote robotics. Both force and vibration information are essential for haptic feedback, and it is important to understand their applicability in different situations. In this study, the relationship between the effective type of haptic feedback and target properties in robot-mediated material discrimination was investigated. A remote-control system including a force presentation device and a wearable vibrotactile display was constructed. In the first experiment, the discrimination performance of material hardness was compared between two types of feedback, force and hybrid (vibrotactile and force) conditions. The results show that both feedback systems allow statistically-significant discrimination of the stimuli, and a significant difference in correct-answer rates between the two feedback conditions was not observed. This indicates that the force system was effective for hardness discrimination, and that there was no superimposed effect of the hybrid system. In the second experiment, the discrimination performance of material roughness was compared between three types of feedback (force, vibrotactile, and hybrid). The results indicate that the rate of correct responses for hybrid feedback condition are significantly higher than those for the force condition. This suggests that hybrid feedback is effective for roughness discrimination. Therefore, the effective type of feedback depends on the properties of target materials, and the superimposed effect of hybrid feedback was only observed in roughness discrimination. These findings play an important role in selecting the best feedback method for a given situation or constructing multiple feedback methods that achieve high discrimination performance.

COVID-19 Detection Using Forced Cough Sounds and Medical Information

The World Health Organization (WHO) has declared the novel coronavirus (COVID-19) outbreak a global pandemic in March 2020. Through a lot of cooperation and the effort of scientists, several vaccines have been created. However, there is no guarantee that the virus will shortly disappear, even if a large part of the population is vaccinated. Therefore, non-invasive methods, with low cost and real-time results, are important to detect infected individuals and enable earlier adequate treatment, in addition to preventing the spread of the virus. An alternative is using forced cough sounds and medical information to distinguish a healthy person from those infected with COVID-19 via artificial intelligence. An additional challenge is the unbalancing of these data, as there are more samples of healthy individuals than contaminated ones. We propose here a Deep Neural Network model to classify people as healthy or sick concerning COVID-19. We used here a model composed by an Convolutional Neural Network and two other Neural Networks with two full-connected layers, each one trained with different data from the same individual. To evaluate the performance of the proposed method, we combined two datasets from the literature: COUGHVID and Coswara. That dataset contains clinical information regarding previous respiratory conditions, symptoms (fever or muscle pain), and a cough record. The results show that our model is simpler (with fewer parameters) than those from the literature and generalizes better the prediction of infected individuals. The proposal presents an average Area Under the ROC Curve (AUC) equal to 0.885 with a confidence interval (0.881 - 0.888), while the literature reports 0.771 with (0.752 - 0.783).

Plastic behavior generation of a robot without force information

Plastic behavior generation of a robot without force information

Development of a Training Simulator for Caregivers' Toothbrushing Skill Using Virtual Reality

Oral care is key to maintaining overall health. Elderly people and people with disabilities who require nursing care can have difficulty in brushing their teeth effectively and may require the assistance of a caregiver. However, to date, an effective method for teaching toothbrushing skills to caregivers and a training system that can be implemented in a non-in-person setting have not been established. Therefore, in this study, we developed a training simulator that enables skill acquisition while learning ideal brushing motion and force information from an arbitrary viewpoint in a virtual reality (VR) space. In this simulator, the position and orientation of the toothbrush as measured by six infrared cameras and the brushing force as measured by a small 6-axis force sensor are displayed in a VR space using computer graphics to match the visual information in the VR space with hand-force information in real space. The experiment was conducted in 10 healthy adults with no specialized skills or knowledge regarding oral care. They were trained to brush the cervical and distal parts of the maxillary central incisors while learning the ideal brushing motion shown by the computer graphics and the appropriate range of brushing force. Afterwards, skills were quantified in terms of brushing motion, brushing force, and plaque removal rate, and the overall effectiveness of the training was evaluated. The results showed that in cervical part brushing, the mean error between the participants' brushing motion and the ideal brushing motion in the brushing direction improved from approximately 4.8 mm before training to 3.9 mm after training. In distal part brushing, the error improved from approximately 3.8 mm to 1.9 mm. Similarly, brushing force improved from approximately 0.8 N before training to the appropriate force of 1.5 N after training for both cervical and distal part brushing. The artificial plaque removal rate improved from 40.3% to 68.7% for cervical part brushing and from 39.5% to 63.3% for distal part brushing, indicating the effectiveness of training using the proposed system. This simulator, which can simultaneously teach both brushing motion and brushing force, is expected to be developed as a new method for teaching toothbrushing skills to caregivers.

A Laminated Continuum Robot With Gripping Force-Sensing Forceps for Robot-Assisted Minimally Invasive Surgeries

Minimally invasive surgeries (MISs) or natural orifice transluminal endoscopic surgeries (NOTESs), such as the transurethral resection of bladder tumor (TURBT), require the surgical robot to be miniaturized to perform surgical procedures in narrow spaces. However, miniaturization of surgical robots poses problems in its fabrication and sensor integration. In this article, a miniature continuum surgical robot is proposed with force-sensing surgical forceps, which can be fabricated through a 2-D lamination process and converted into 3-D through folding. This multimaterial laminated structure facilitates the integration of photoelectric sensors into the surgical forceps. Then, the gripping force of the surgical forceps is sensed from the change of the output voltage of the integrated photoelectric sensor. The sensitivity of gripping force to the change of output voltage is 3.6 N/V, and the maximum error during calibration is 0.03 N. The proposed robot and the gripping force sensing method are validated through simulations and experiments, such as gripping and releasing objects of various hardness and heights. It is also validated that the gripping force information could also improve the gripping performance in the object handling experiment and ex vivo porcine intestine gripping experiment.

License Plate Recognition via Attention Mechanism

License plate recognition technology use widely in intelligent traffic management and control. Researchers have been committed to improving the speed and accuracy of license plate recognition for nearly 30 years. This paper is the first to propose combining the attention mechanism with YOLO-v5 and LPRnet to construct a new license plate recognition model (LPR-CBAM-Net). Through the attention mechanism CBAM (Convolutional Block Attention Module), the importance of different feature channels in license plate recognition can be re-calibrated to obtain proper attention to features. Force information to achieve the purpose of improving recognition speed and accuracy. Experimental results show that the model construction method is superior in speed and accuracy to traditional license plate recognition algorithms. The accuracy of the recognition model of the CBAM model is increased by two percentage points to 97.2%, and the size of the constructed model is only 1.8 M, which can meet the requirements of real-time execution of embedded low-power devices. The codes for training and evaluating LPR-CBAM-Net are available under the open-source MIT License at: .

A Novel Fault Detection and Isolation Technology for Redundant Inertial Measurement Unit

The fault detection and isolation (FDI) accuracy directly affects the redundant inertial measurement unit (RIMU) working state. The previous methods are difficult to apply to the four-axis system because they do not have enough information to allow fault isolation. Therefore, in this paper, an extended RIMU system is proposed to implement fault location in a four-axis system. The extended system consists of three additional accelerometers and a four-axis RIMU. The specific force information in the gyro-free strapdown inertial navigation system (GFSINS) is used to obtain reference angular velocity information. Then, based on the extended system, a fault detection procedure is designed to achieve faulty gyro isolation and system reconfiguration. Simulation results show that the extended system can effectively isolate gyroscope failures in a four-axis RIMU system. Compared to the non-redundant three-axis system, the extended system is 1.75 times more reliable and has 37% better average navigation accuracy.

Effects of Feedback of Fingertip Force Information with Temporal Coded Vibration Stimulation on Precision Grasping Tasks

Tactile information plays an important role in human manipulation of objects; however, prosthetic limb placement or teleoperation requires the manipulation of alternative bodies in the absence of tactile sensations. To provide an alternative to tactile sensation, this study proposes and assesses a continuous feedback scheme with temporally coded vibration. This scheme was designed to provide discrete intended tactile information in response to changing object-controlled situations by repetitively presenting time-coded vibration patterns. The effects of the proposed scheme on an object with acatch-and-hold task in virtual reality were confirmed. Compared to the control feedback scheme that provides vibration only when the balance of the virtual grip force and object position is changed, the proposed feedback scheme has a better effect in terms of the success rate of holding on to the object with an appropriate holding force during the task. The effect is larger, especially in the invisible task condition, suggesting that the increased amount of information with coded vibration patterns can be used without any special training, especially without visual information. Considering the existing studies that show the effect of a feedback scheme in response to motion events, the continuous feedback scheme proposed in this study may be more suitable for movements that require sequential coordination and passive responses than stimulation methods based on motion events. This feedback scheme has potential applications not only in tele-technology but also in healthcare, such as rehabilitation.

Optical Tweezers to Force Information out of Biological and Synthetic Systems One Molecule at a Time

Over the last few decades, in vitro single-molecule manipulation techniques have enabled the use of force and displacement as controlled variables in biochemistry. Measuring the effect of mechanical force on the real-time kinetics of a biological process gives us access to the rates, equilibrium constants and free-energy landscapes of the mechanical steps of the reaction; this information is not accessible by ensemble assays. Optical tweezers are the current method of choice in single-molecule manipulation due to their versatility, high force and spatial and temporal resolutions. The aim of this review is to describe the contributions of our lab in the single-molecule manipulation field. We present here several optical tweezers assays refined in our laboratory to probe the dynamics and mechano-chemical properties of biological molecular motors and synthetic molecular devices at the single-molecule level.

Effects of movement direction and limb dominance on ankle muscular force in sidestep cutting.

Sidestep cutting is a critical movement in sports. However, biomechanical research on sidestep cutting has not hitherto reached a consensus. In order to investigate the effects of limb dominance and movement direction on ankle and subtalar joints during sidestep cutting, twelve physically active male participants were recruited in the present study. Trajectory and ground reaction force data were collected by the motion capture system and force platform. Kinematics, kinetics, and muscle forces information were obtained by running OpenSim. Two-way repeated measures ANOVA was performed with movement direction and limb dominance as independent variables. We found that movement direction had a significant effect on ankle dorsiflexion angle. In contrast, the factor of limb dominance had no effect on ankle and subtalar joints angles. For ankle joint moment, the plantarflexion moment was greater by performing a 45° sidestep cutting or using the dominant limb, while the subtalar joint moment was not affected by these two variables. In terms of muscle forces, the soleus of the dominant limb generated greater plantarflexion muscle force on the sagittal plane, while the non-dominant limb tended to contract more strongly (peroneus longus and peroneus brevis) on the frontal plane to stabilize the subtalar joint. Meanwhile, a smaller sidestep cutting angle made participants generate greater plantarflexion muscle forces (soleus and gastrocnemius). In conclusion, our findings indicated that participants should take limb dominance and movement direction into consideration for enhancing athletic performance and reducing the risk of injury during sidestep cutting.

Role of path information in visual perception of joint stiffness.

Humans have an astonishing ability to extract hidden information from the movement of others. In previous work, subjects observed the motion of a simulated stick-figure, two-link planar arm and estimated its stiffness. Fundamentally, stiffness is the relation between force and displacement. Given that subjects were unable to physically interact with the simulated arm, they were forced to make their estimates solely based on observed kinematic information. Remarkably, subjects were able to correctly correlate their stiffness estimates with changes in the simulated stiffness, despite the lack of force information. We hypothesized that subjects were only able to do this because the controller used to produce the simulated arm's movement, composed of oscillatory motions driving mechanical impedances, resembled the controller humans use to produce their own movement. However, it is still unknown what motion features subjects used to estimate stiffness. Human motion exhibits systematic velocity-curvature patterns, and it has previously been shown that these patterns play an important role in perceiving and interpreting motion. Thus, we hypothesized that manipulating the velocity profile should affect subjects' ability to estimate stiffness. To test this, we changed the velocity profile of the simulated two-link planar arm while keeping the simulated joint paths the same. Even with manipulated velocity signals, subjects were still able to estimate changes in simulated joint stiffness. However, when subjects were shown the same simulated path with different velocity profiles, they perceived motions that followed a veridical velocity profile to be less stiff than that of a non-veridical profile. These results suggest that path information (displacement) predominates over temporal information (velocity) when humans use visual observation to estimate stiffness.