Research on Sliding Perception of Robotic Hands Based on FBG Sensing
As an important carrier of artificial intelligence,robots exhibit many abilities that humans do not possess.Robotic hands are an important component of robots and an important subfield of the intelligent robotics industry.Robotic hands surpass human hands in accuracy and stability.It can perform precise welding and assembly processes in the production workshop,replace humans in completing dangerous tasks such as explosive dismantling,and can also be used for deep-sea and aerospace exploration.However,traditional robotic hands are purely functional and lack the ability to perceive the surrounding environment,making it difficult to respond promptly to changes in the working environment.The application of sensor technology is a prerequisite for achieving robot control,interaction,and intelligence.The integration of sensor technology is crucial for robot control,human-machine interaction,and intelligence.Sending sensor exported data to the feedback control system improves the operational efficiency of the robotic hand.Electrical sensors have been widely used in other fields.Although the technology is mature and capable of sensing physical quantities such as force and temperature,but due to their poor flexibility,susceptibility to electromagnetic interference,difficulty in wiring,and poor flexibility,they can not fully meet the needs of intelligent sensing for robotic hands.FBG sensors use wavelength encoding,so the energy loss and electromagnetic field changes of the optical fiber can not interfere with the detection results.They also have the advantages of high accuracy,fast response speed,and low loss.FBG sensors exhibit excellent flexibility and scalability,with minimal impact on the maneuverability and agility of robotic hands.Therefore,FBG sensors are an ideal choice for sensor applications in robotic hands.In this paper,a flexible sliding sensor is proposed for detecting the sliding direction and speed of robotic hands.The sensor uses silicone rubber as the packaging material,which can improve sensitivity of the sensor and protect the FBG.Firstly,FEA was conducted using ANSYS software to obtain theoretical values of strain changes over time at different positions during the sliding process.Then,based on the simulation results of FEA,a sliding sensor was designed and the sensor fabrication was completed.Finally,a sliding detection platform was built to experimentally verify the simulation results.The experimental setup includes an FBG demodulator,sensors,PC,motor,weights,and pulleys.The motor is used to horizontally and uniformly pull weight through the surface of the sensor.The experimental results show that the sensitivity of each FBG sensor encapsulated in silicone rubber has sensitivity to pressure of 9.26,7.83 and 7.76 nm/MPa,and the sensitivity to shear force is 1.74,3.17 and 3.29 nm/MPa,respectively.When the slider slides over the sensor surface in different directions,the wavelength shift curve of each FBG shows different trends of variation.By using SVM model and 1D-CNN network to identify the sliding characteristics,the accuracy of sliding direction judgment reaches 91.0%and 93.5%,respectively.The sliding time of the slider is obtained from the wavelength shift curve,and the sliding speed can be obtained by dividing the sliding distance by the sliding time.The maximum relative error of sliding speed calculation is 9.29%.The research results of this paper have certain practical value for the research and promotion of robotic biomimetic skin.
NETL
NSTL
万方数据
