• Journal of the Korean Society for Precision Engineering
• /
• v.26 no.8
• /
• pp.132-141
• /
• 2009

In this paper, we describe finger force estimation from surface electromyogram (sEMG) data for intuitive and delicate force control of robotic devices such as exoskeletons and robotic prostheses. Four myoelectric sites on the skin were found to offer favorable sEMG recording conditions. An artificial neural network (ANN) was implemented to map the sEMG to the force, and its structure was optimized to avoid both under- and over-fitting problems. The resulting network was tested using recorded sEMG signals from the selected myoelectric sites of three subjects in real-time. In addition, we discussed performance of force estimation results related to the length of the muscles. This work may prove useful in relaying natural and delicate commands to artificial devices that may be attached to the human body or deployed remotely.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.59 no.5
• /
• pp.93-99
• /
• 2017

A model was developed using Artificial Neural Networks (ANNs) based on Principal Component Analysis (PCA), to accurately predict the air humidity inside an experimental greenhouse located in Daegu (latitude $35.53^{\circ}N$, longitude $128.36^{\circ}E$, and altitude 48 m), South Korea. The weather parameters, air temperature, relative humidity, solar radiation, and carbon dioxide inside and outside the greenhouse were monitored and measured by mounted sensors. Through the PCA of the data samples, three main components were used as the input data, and the measured inside humidity was used as the output data for the ALYUDA forecaster software of the ANN model. The Nash-Sutcliff Model Efficiency Coefficient (NSE) was used to analyze the difference between the experimental and the simulated results, in order to determine the predictive power of the ANN software. The results obtained revealed the variables that affect the inside air humidity through a sensitivity analysis graph. The measured humidity agreed well with the predicted humidity, which signifies that the model has a very high accuracy and can be used for predictions based on the computed $R^2$ and NSE values for the training and validation samples.

• Journal of the Korean Society for Precision Engineering
• /
• v.26 no.11
• /
• pp.99-107
• /
• 2009

Since a long work piece influences the natural frequency of the entire system with a miniature high speed spindle, a bar-feeder is used for a long work piece to improve the vibration characteristics of a spindle system. Therefore, it is very important to design optimally support positions between a bar-feeder and a long work piece for a miniature high speed spindle system. The goal of the current paper is to present an optimization method for the design of support positions between a bar-feeder and a long work piece. This optimization method is effectively composed of the method of design of experiment (DOE), the artificial neural network (ANN) and the genetic algorithm (GA). First, finite element models which include a high speed spindle, a long work piece and the support conditions of a bar-feeder were generated from the orthogonal array of the DOE method, and then the results of natural vibration analysis using FEM were provided for the learning inputs of the neural network. Finally, the design of bar-feeder support positions was optimized by the genetic algorithm method using the neural network approximations.