Session: 12-10-03: General: Mechanics of Solids, Structures and Fluids

Paper Number: 116869

116869 - In-Situ Calibration for Load Cells in 3d Printed Bipedal Robot Using 3D Modeling in Computer-Aided Design Environment 

Load cell or force sensor in general is very important in robotic, especially for bipedal robot when its control system need the force reaction reading in the robot structure as the feedback when the robot is standing and moving. However, after experience difference force overtime, there will be some plastic deformation and fatigue on the load cell material which cause some discrepancy between the real force applied on the load cell and its force reading. Therefore, it is typically for the load cell to be recalibrated annually to maintain its accuracy. In some cases, the operators even need to recalibrate the load cell whenever there are some changes in the force reading system. One of the most popular methods to do load cell calibration is the operator takes the load cell out of the host system, puts known difference weights on top or hang weights below the load cell. Then the operator uses the raw values output from the load cell corresponding to the weights attach with it to get the conversion equation from the raw reading value to force value. The advantage of this method are high accuracy and it does not require a lot training to do the calibration. However, the biggest drawback is since the operator need to take the load cell out from the host system, it will take a lot of time and labor work if the host system is very complicated and not maintenance friendly. Before this, some researchers did some work with the in-situ calibration for the 6-axis force sensor. In these methods, the operator controls the robot arm to move into difference posture and collect the raw values from the force sensor. Then they use matrix transformation to calculate the force that applied to the force sensor. These methods let the operator to calibrate the force sensor without taking the sensor out of the host system. However, these methods only work with force sensor attached to the end effector of the robotic arm, and they need industrial robotic arm with very high precision control to do the calibration. Therefore, they are incompatible with robots that has 3D printed structure since this type of robot have low precision in its control and high tolerance in its structure.    

In this paper, the authors will present a difference approach to do load cell in-situ calibration that can work with load cells attached to difference positions of the robot and also compatible with 3D printed robots. In this method, the operator only needs 3D part models of the robot in computer-aided design (CAD) environment and conduct the static simulation in CAD environment based on these part models. To reduce the part complexity so a computer with limited computing power can do the simulation, the authors will present their methods to simplified the geometry of the 3D model parts. This simplification method will make 3D model parts to be able to represent the physical properties of their own parts in real life while the simulation does not require a lot of computing power. In the end of this paper, the authors will present the experiments they conducted and the results from the experiment. Based on the results, the accuracy of this in-situ calibration method is similar to the common load cell calibration method which the error percentage comparing to the whole range of load cell measurement from -2224N to 2224N is 0.157%.

Presenting Author: Tung Le Virginia Polytechnic Institute and State University

Presenting Author Biography: Tung Xuan Le is the 2nd year Master of Science student Mechanical Engineering department of Virginia Tech. He received B.Sc Degree of Mechanical Engineering from Virginia Tech in 2021. He is currently working in Terrestrial Robotics Engineering and Controls Lab (TREC Lab) for his Master thesis under the advising of Dr. Alexander Leonessa.

Authors:

Xuan Tung Le Virginia Polytechnic Institute and State University
Connor Herron Virginia Polytechnic Institute and State University
Alexander Leonessa Virginia Polytechnic Institute and State University