Force/Moment Isotropy of 8/4-4 Parallel Six-Axis Force Sensor Based on Performance Atlases

A six-axis force sensor with parallel 8/4-4 structure is introduced and its measurement principle is analyzed.Based on condition numbers of Jacobian matrix spectral norm of the sensor,the relationship between the force and moment isotropy and some structural parameters is deduced.Orthogonal test methods are used to determine the degree of primary and secondary factors that have significant effect on sensor characteristics.Furthermore,the relationship between each performance index and the structural parameters of the sensor is analyzed by the method of the atlas,which lays a foundation for structural optimization design of the force sensor.

Dynamic analysis of double-layer and pre-stressed multi-limb six-axis force sensor

In order to adapt to the specific task, the six-axis dynamic contact force between end-effectors of intelligent robots and working condition needs to be perceived. Therefore, the dynamic property of six-axis force sensor which is installed on the end-effectors of intelligent robots will have influence on the veracity of detection and judgment to working environment contact force by intelligent robots directly. In this paper, dynamic analysis to double-layer and pre-stressed multi-limb six-axis force sensor is conducted. First, the structure of the sensor is introduced, and the limb number is confirmed by introducing the related definitions of convex analysis. Then, based on vibration of multiple-degree-of-freedom system, a mechanical vibration simplified model of double-layer and pre-stressed multiple limb six-axis force sensor is set up. After that, movement differential equations of sensor and the response of analytical expression are deduced, and the movement differential equations is solved. Finally, taking the double-layer and pre-stressed seven limb six-axis force sensor as an example, numerical calculation and simulation of deriving result is conducted, which verify the correctness and feasibility of the theoretical analysis.

Optimal design and experiment research of an orthogonal-parallel six-axis force/torque sensor

A novel orthogonal-parallel six-axis force/torque sensor is studied based on a modified Stewart platform architecture,and the optimal design and experiment research of the sensor are discussed.Firstly,the model of orthogonal parallel six-axis force/torque sensor based on improved Stewart platform architecture and its static mathematical model are proposed.Secondly,according to the actual working condition of the sensor,the sensor is optimized and the optimal solution is obtained.Then,the experimental prototype and calibration system is developed.Finally,the superiority of the sensor structure and the effectiveness of the optimization method are verified by calibration experiments.The results of the proposed method are useful for the further research and application of the orthogonal-parallel six-axis force/torque sensor.

Feature Fusion-Based Deep Learning Network to Recognize Table Tennis Actions

A system for classifying four basic table tennis strokes using wearable devices and deep learning networks is proposed in this study.The wearable device consisted of a six-axis sensor,Raspberry Pi 3,and a power bank.Multiple kernel sizes were used in convolutional neural network(CNN)to evaluate their performance for extracting features.Moreover,a multiscale CNN with two kernel sizes was used to perform feature fusion at different scales in a concatenated manner.The CNN achieved recognition of the four table tennis strokes.Experimental data were obtained from20 research participants who wore sensors on the back of their hands while performing the four table tennis strokes in a laboratory environment.The data were collected to verify the performance of the proposed models for wearable devices.Finally,the sensor and multi-scale CNN designed in this study achieved accuracy and F1 scores of 99.58%and 99.16%,respectively,for the four strokes.The accuracy for five-fold cross validation was 99.87%.This result also shows that the multi-scale convolutional neural network has better robustness after fivefold cross validation.