Traditional Adaptive Control on Robot Manipulator ( 2 link)

on

AIM

In this project we design a controller, check its performance, compare and contrast the results from the simulations on a Robot Manipulator.





There are 6 simulations done with 6 controllers on a two-link rigid revolute robot manipulator with the following dynamics

  • Simulation 1: Traditional and Composite Adaptive controller
  • Simulation 2: Composite adaptive controller with RISE component
  • Simulation 3: adaptive controller with CL and ICL update law
  • Simulation 4: RISE based modular adaptive controller
  • Simulation 5: Repetitive learning controller
  • Simulation 6: Neural Network-based controller with a discontinuous sliding-mode feedback control law for the dynamics Neural Network-based controller with a continuous RISE feed-back control law for the dynamics

Control Designs and update law designs are clearly mentioned in this repo GitHub Files

 Simulation 1: Traditional and Composite Adaptive controller

Traditional adaptive controller
Traditional adaptive controller

Composite adaptive with gradient update law

Composite adaptive with least squares update law 


 We don't see a lot of variation in the performance, however this controller is only feasible when we know the dynamics of the system and we can bear slight oscillations in error. 

To be Continued

Comments

Post a Comment