Control of End-Effector of a Multi-link Robot with Joint and Link Flexibility

Abstract

Flexibility in manipulators/robots is due to both joint and link flexibility that makes up the system. Flexible robots are preferred over conventional rigid robots in applications like invasive surgeries, space applications, and industries due to their prompt response, low energy requirement, faster operational speeds, and low weight to power ratio. Due to inherent flexibility, accurate positioning of end-effector in required path is difficult. Moreover flexibility of link makes it an infinite degree freedom system and mathematics is very involved. To simplify the problem and get reasonable results, flexible links are modeled based on Euler–Bernoulli beam theory and Assumed mode method is implemented. Joint flexibility is because of small clearances that are inherently present in the joint, because of both manufacturing and assembling constraints, these clearances cause sudden impacts between the joining parts (journal and bearing) resulting in impact force generation as the joints are manipulated. Resulting impact (hertzian contact) forces increase the overall input torque required to manipulate the end-effector according to our wish. This paper’s objective is to build a dynamic model of a two-link RR type planar manipulator with link and joint flexibility, and determine the maximum error of tip position between a robot with/without flexibility, as the end effect or travels in required vertical path with payload. Further, apply orthodox control strategies (PD, PI, and PID) to reduce the error. The end-effector carries a payload equals its links mass. Using MSC Adams and MATLAB softwares, a co-simulation approach is developed. Both the controllers (PI, PID) radically reduced error through several iterations, PID control strategy achieved better results than PI controller and by both approaches, more than 60% of the positional error is reduced.