علي پارسايي كيا

Collaborating mechanical arms are recognized as one of the most widely used robots in the industry and medical science, the most important applications of which are humanoid robots, surgeon arms and arms used in the automotive industry. Due to the Industrial development and the need for higher accuracy to achieve the design goal, these types of robots have attracted more and more researchers. For this reason, this dissertation addresses the challenges of these types of systems to improve their performance and use more than before. Among the most important challenges in partner mechanical arms are the optimal external load distribution between the arms, the uncertainties in the modeled system, and the design of a control algorithm to better control these types of arms. The optimal distribution of external load has a significant effect on factors such as the amount of power consumption of the joints, the maximum load that can be carried by each arm and also the amount of toe error. Therefore, in order to optimally distribute the external load, a new method called the optimal load distribution method based on the saturation limit of the actuators (motor) of each joint (SPLLDA) has been introduced and proposed. This method, according to the tolerable capacity of each joint actuator, according to their saturation limit, optimally distributes the external load between the cooperating arms at any time so that the design goal can be optimally pursued. The following are three performance indicators to eva‎luate the optimal external load distribution, which are about twice the indicators of the previous load distribution methods for the proposed method. A sliding mode controller with a function that is resistant to uncertainties is used to track the path and adjust the state variables. But to compensate for the greater uncertainties and disturbances of the system, a neural network compensator is designed to determine the optimal control weights in this controller by this trained network. The method is that the optimal instantaneous control weights according to the amount of error and power consumption of the joints at any time are extracted by a trained neural network. Finally, to validate the proposed method, simulations and practical experiments are performed on the cooperating arms of three degrees of freedom, which show the results of improving the performance of the intelligent sliding mode controller compared to the usual mode. Thus, the intelligent sliding mode controller was 2.7 times in the control performance index and 1.2 times in the error index in comparison with the usual method of the sliding mode controller.

بازوهاي مكانيكي همكار , كنترل‌كننده مود لغزشي (SMC) , توزيع بهينه بار مبتني بر حد اشباع موتورها (SPLLDA) , جبران‌ساز شبكه عصبي , كنترل‌كننده هوشمند مقاوم

Cooperative Manipulators , NN-SMC Controller , Neural Network Compensator , ISPLLDA

ali parsaie kia

moharam habibnejad kurayem