Design and Simulation of an Adaptive Neuro_ controller for a Wire Driven Flexible Arm Robot

Nowadays, robots are employed widely, and they are utilized in many operations and occupations. They are not just luxury devices or entertainment tools, but they are used to reduce costs and increase the accuracy of functions such as robotic surgery or agricultural harvesting in countries with high labor costs. One type of robot is a wire-driven robot developed recently for industrial inspections, laparoscopic surgeries, and crop harvesting. This type of robot has a limited range of payloads supporting and accuracy in positioning, but simple structure and adaptability by working space lead to focus by researchers. The current paper introduces a wire-driven flexible robot where its kinematic and dynamic models are presented based on the Euler-Bernoulli beam theory. Regards to stated models, a proportional-derivative-integrator controller (PID) is employed for this dynamic system. In addition, an adaptive neuro controller with two hidden layers and five neurons was established for that. Various functions such as step, ramp, sinusoidal, random, and square signals are exposed to both controllers as inputs to study their stability and accuracy. The adaptive neuro controller responses are as good as the PID controller though in the face of random and sinusoidal inputs did not show better performance.