The Operation Intention Recognition Algorithm For Transfer Robot Research And Implementation

The rapid development of intelligent manufacturing,human-robot collaboration is getting more and more attention.In the process of human-robot cooperative transferring,human and robot work in the same space,not only to ensure the stability of the robot,accurate,fast,but also to ensure the safety of the human,it's essential the robot to identify the intentions of people.At present,domestic and foreign researchers have done a lot of research on robot intention recognition in the process of human-robot collaboration,but there are still many shortcomings,such as,contact force intention recognition requires long time contact and low rate of speech commands intention recognition.In view of the above problems,this paper mainly does the following work:First,using Quanser software and hardware of Canada and Panasonic servo system,a one-dimensional planar transferring robot hardware-in-loop simulation system is established.The FSR SMD resistor type thin film pressure sensor with its processing circuit is installed in the end of the robotic arm,which enables the system to recognize the operator's contact force intention.Meanwhile,a voice processing system based on STM32 is installed to enable the system to recognize the operator's speech intention.Second,intention recognition is carried out in time domain and frequency domain.With one-dimensional force signals as intention recognition signal,FSR SMD resistor type thin film pressure sensor is used to collect a brief contact force,the application of the given classification rules,the robot to the operator intention recognition is realized,in the case of without the need for constant contact,the robot automatically reaches the intended target position of the operator.The intent recognition control algorithm is verified on a one-dimensional planar transferring robot simulation system based on contact force intention recognition.The results show that in the case of forward contact or reverse contact,the robot can reach the target position of the operator accurately and quickly,indicating that this method can complete the human-robot cooperative transferring task based on the contact force transferring robot.Thirdly,in order to solve the problem that the speech signal can not be recognized effectively by the traditional processing method.In this paper,the deep learning method is used to process speech signal for realizing intention recognition,speech signal processing with the traditional time domain or frequency domain method the effect is not ideal,in this paper,the convolutional neural network is used to train the instruction data set of Google on PC,the trained network is used to recognize speech commands,using STM32 as communication between one-dimensional planar transferring robot system and PC,using four fixed movement Angle,two direction,that is,"Left" and "Right" are used as direction commands,and "One","Two","Three" and "Four" are used as four fixed motion angles.The results show that the robot can recognize the speech command of the operator,that is,the intention of the operator,and make corresponding actions.Fourth,in order to further improve the transferring robot to speech recognition accuracy of this kind of signal,this paper puts forward using residual network speech instruction set for training,the trained network is used to analyse the speech recognition,the results show that the relative convolution network method improves the accuracy of about 3%,the recognition effect is also verified by using the one-dimensional planar transferring robot hardware-in-the-loop simulation system platform with speech processing system,the results show that the transferring robot intention recognition can be performed.In this paper,the traditional method and the deep learning method are used to carry out the intention recognition of the handling robot,and the one-dimensional planar hardware-in-the-loop simulation system is built to verify the algorithm.The experiment proves the effectiveness of the algorithm,and provides a feasible scheme for the transferring robot's intention recognition of force and speech signal.