To have the best performance (trajectory follow, disturbance rejection, torque performance and error performance), which type of controller do you recommended?
Linear or nonlinear controller?
Conventional or intelligent controller?
I am enthusiastically looking forward to hearing from you about this question.
Regards
I recommend - Intelligent Nonlinear Controller!
Best wishes
Sundar
Dear Prof.,
I'd like to know about the reason of this selection.
Regards
Dear Farzin:
As the dynamics of a robot manipulator is nonlinear, your question can be even generalized into what is the best control method for the design (output regulation, disturbance rejection, robustness, optimization) of a nonlinear control system.
As Chris Byrnes would say - "Take the bull by the horns" - so we go for a nonlinear controller to get satisfactory global results instead of a linear controller which may give good local results. Enhancing the nonlinear controller with intelligent control (fuzzy logic control, neuro-fuzzy control or some tuning using evolutionary algorithms) is known to have given better results. These are general remarks, but there are many comparative analysis done using linear controllers, nonlinear controllers & neuro-fuzzy or GA/PSO/ACO-enhanced nonlinear controllers etc.
Best wishes
Sundar
Dear Sundar
Thank's for your answer. I'd like to know about Intelligent conventional controller such as (Fuzzy sliding mode controller compare with Pure fuzzy logic controller). What is your idea about it? which method is better pure intelligent control or intelligent conventional control?
Regards
Farzin Piltan
A biological perspective may be interesting as well.
The best control POSITION for a robot would be influenced by what controller is chosen, and, as the biologist would perhaps say, by the specific task that the robot is engaged in.
Similarly, the type of controller could be chosen by the kind of task one wishes to accomplish.
Biological systems tend to be redundant and lazy. This general observation is reflected in 'the uncontrolled manifold hypothesis' (Schoner and Scholz) and the 'minimum intervention principle' (Todorov and Jordon). There are controllers designed on these principles. Briefly: if a perturbation to the system (robot) displaces the system in a direction that does not affect the task, then that perturbation and its effect are ignored by the controller.
"If it aint broken, dont fix it".
Dear Farzin:
Dynamic Adaptive Backstepping procedure can also be tried. This method caters for matching as well as unmatching uncertainties which makes system robust. Tracking and regulation problems can be handled nicely. This method involves extensive mathematical calculations. Another robust technique is integral sliding mode which takes care of chattering which is a shortcoming of conventional sliding mode.
Mahmood:
Thanks for your reply. Yes, integral sliding mode control is very popular these days, and recently I read a very good paper by a friend, Dr. M. Feki in an IJMIC issue (2014).
Best wishes,
Sundar
Dr Mahmood, and Dr Sunder
Thank's for your reply. The backstepping coefficient optimization is very difficult compare to SMC. What is your comments about it?
I design three types of controller for robot manipulator: Computed torque control, sliding mode control and backstepping control, but the best result is about CTC, however SMC and backstepping is more robust that CTC. Do you have any comments?
Best Regards
Dear Dr. Farzin Piltan
I agree with your first comment, but can not comment on your second observation, bc, have not yet applied CTC.
To my best knowledge : Flatness-based control and it refinement Exactfeedforward linearization
Article Robustness Analysis With Respect to Exogenous Perturbations ...
Hi dear Dr. Farzin Piltan
An recent and effective control technique is so called Active Disturbance Rejection Control (ADRC). This approach is due to professor J. Q. Han.
Conference Paper Active Disturbance Rejection Control: A paradigm shift in fe...
Article Active disturbance rejection control for trajectory tracking...
Article Linear active disturbance rejection control of underactuated...
We have been using the DAQ2GO as a controller and data taker for a number of custom systems. It has interfaces for analog, digital, frequency, current loop, etc. It programs in Excel and communicates to all these ports in easily programmed ways. It can also communicate with RS485 and RS232 devices. Uses all the features and power of Microsoft office. Works with 32 and 64 bit machines. Any win from xp forward, and office product 2007 forward.
Please refer to these articles:
https://www.researchgate.net/publication/313576646_From_PID_to_Nonlinear_State_Error_Feedback_Controller
https://www.researchgate.net/publication/312046377_Improved_Sliding_Mode_Nonlinear_Extended_State_Observer_based_Active_Disturbance_Rejection_Control_for_Uncertain_Systems_with_Unknown_Total_Disturbance
https://www.researchgate.net/publication/309655805_On_the_Improved_Nonlinear_Tracking_Differentiator_based_Nonlinear_PID_Controller_Design
Article From PID to Nonlinear State Error Feedback Controller
Article Improved Sliding Mode Nonlinear Extended State Observer base...
Article On the Improved Nonlinear Tracking Differentiator based Nonl...
We have been studying various control algorithms and types of actuators form many applications. Luckily, we have the Excel based system with high performance electronics (DAQ2GO) so it's really easy to try out various algorithms and simultaneously measure how well they work. What has impressed us most is the woeful inadequacy of regular pid controllers for doing things easily, efficiently and well. High speed, computerized motion control with position feedback seems to work very well for many applications.
Please refer to this article:
https://www.researchgate.net/publication/318726178_An_Improved_Active_Disturbance_Rejection_Control_for_a_Differential_Drive_Mobile_Robot_with_Mismatched_Disturbances_and_Uncertainties
Conference Paper An Improved Active Disturbance Rejection Control for a Diffe...
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory