EE 7280 Project # 1 (Part B)
Intelligent Control of Robotic Systems: Fault Accommodation
Read the paper by Vemuri and Polycarpou. Develop Matlab/Simulink model to implement the
following:
1. Assume the modeling uncertainty is given by
F = ” θ θ1 1q q__1 2 #
τd = ” θ θ2 2 sin(30 sin(30t t) ) # ;
where the unknown parameters θ1 2 [0 0:5], θ2 2 [0 0:5] (In your simulation of robotic
dynamics, use the following values θ1 = 0:45 and θ2 = 0:4). Derive a bound function on the
modeling uncertainty and design an adaptive threshold for fault detection (Note that the
actual values of θ1 and θ2 ar not available for threshold design).
2. Consider a fault which occurs at t = 10 s and manifest itself as a nonlinear change in the
robotic system dynamics described by
f(x; u) = ” 2:5 ∗ [37:5(q1)2 + 50( 0 q1)2(q2)2 + 3:5q2] # :
Demonstrate the fault detection performance of your algorithm designed in Step 1. Additionally, develop a dynamic estimation model using a Sigmoidal neural network (SNN). Design
adaptive laws for the SNN in the estimation model to learn the unknown fault function f(x; u)
after fault detection.
3. Use the fault approximation model provided by the SNN to design an intelligent controller
to automatically accommodate the effect of the unknown fault after fault detection.
In your report, you should show all details of the above design and simulation results illustrating
the following:
• Learning capability of the neural network by comparing the neural network output with the
actual fault function (see, for instance, Fig. 2 of the paper).
• Advantages of the intelligent control method by comparing the control performances of the
nominal controller (designed in project 1) with the intelligent controller in the presence of
the fault.
AssignmentTutorOnline
- Assignment status: Already Solved By Our Experts
- (USA, AUS, UK & CA PhD. Writers)
- CLICK HERE TO GET A PROFESSIONAL WRITER TO WORK ON THIS PAPER AND OTHER SIMILAR PAPERS, GET A NON PLAGIARIZED PAPER FROM OUR EXPERTS
