The minimum gripping force applied is expected to prevent objects from mechanical damage when an agricultural robot is applied to handle and manipulate fruits and vegetables.In this research,a sensitive slipping senso...The minimum gripping force applied is expected to prevent objects from mechanical damage when an agricultural robot is applied to handle and manipulate fruits and vegetables.In this research,a sensitive slipping sensor was developed with a piezo resistor to control the griping force of the agricultural robot.Firstly,an output of the slipping sensor was analyzed in a frequency domain by using a short time Fourier transform.Then rules for discriminating slipping signal from the output of a slipping sensor were proposed based on detail coefficients of discrete wavelet transform.Finally,a controller based on adaptive Neuro-Fuzzy inference system was developed to adjust the grasping force of the agricultural robot in real time.The detail coefficients and the normal gripping force were applied as input of the controller,and Fuzzy rules were simplified through subtractive clustering.With a two-finger end-effector of the agricultural robot,the experimental results showed that the slipping signal could be effectively extracted regardless of change in the normal gripping force,and the gripping force had been controlled successfully when grasping tomatoes and apples.This method was a promising way to optimize the gripping force of the agricultural robot grasping the fruits and vegetables.展开更多
基金This work was supported by the Key Research and Development Program of Jiangsu(Grant No.BE2017370)the National Natural Science Foundation of China(Grant No.31471419)the Natural Science Funds of Jiangsu(Grant No.BK20140729).
文摘The minimum gripping force applied is expected to prevent objects from mechanical damage when an agricultural robot is applied to handle and manipulate fruits and vegetables.In this research,a sensitive slipping sensor was developed with a piezo resistor to control the griping force of the agricultural robot.Firstly,an output of the slipping sensor was analyzed in a frequency domain by using a short time Fourier transform.Then rules for discriminating slipping signal from the output of a slipping sensor were proposed based on detail coefficients of discrete wavelet transform.Finally,a controller based on adaptive Neuro-Fuzzy inference system was developed to adjust the grasping force of the agricultural robot in real time.The detail coefficients and the normal gripping force were applied as input of the controller,and Fuzzy rules were simplified through subtractive clustering.With a two-finger end-effector of the agricultural robot,the experimental results showed that the slipping signal could be effectively extracted regardless of change in the normal gripping force,and the gripping force had been controlled successfully when grasping tomatoes and apples.This method was a promising way to optimize the gripping force of the agricultural robot grasping the fruits and vegetables.
