Controller and Observer
Design
(Design of Control System in State Space)
(Design of Control System in State Space by Pole
placement)
References
• Dr. Radhakant Padhi, Asstt. Prof, IISC, Bangalore, through NPTEL
• Modern Control Engineering by Katsuhiko Ogata, PHI Pvt. Ltd New
Delhi
Pole Placement
Controller Design
Pole Placement Technique
• Poles of a control system (stable/unstable) can be place at desired
location by pole placement technique. This is done to
• Improve the performance of the system
• Make the system stable
• Increase the damping
• Increase the response time
• Etc
Pole Placement Technique
• Assumptions are
1. The system is completely state controllable
2. The sate variable are measureable and available for feedback
3. Control input (u) is unconstrained and single
Note: For multi input system, the state feedback gain matrix is not unique
Pole Placement Technique
• Objective:
• The closed loop poles should lie ,…. Which are their “desired locations”.
• Difference from classical approach:
• Not only the dominants poles, but “all poles” are forced to lie at specified
desired locations.
• In classical approach only dominants poles are placed at desired location
• Necessary and Sufficient condition:
• The system is completely state controllable
Philosophy of Pole placement control
design
• Let a system is represented by
• ---(1)
• Put input u as
• , put in equation (1)
• K is called state feedback gain matrix (1xn) and X is state vector (nx1)
• So KX will be scalar (=> single input)
• ---(2)
• New closed loop state transition matrix
• Its time response
• ---(3)
Philosophy of Pole placement control
design …
˙
𝑋
∫
X
˙
𝑋
∫
X B ++
B +
u +
u A
A
Fig 1: Open loop Control system -K
Fig 2: Closed loop Control system
With u=-KX
• Philosophy: The matrix K is designed such a way that the two
characterize equations are having same poles
Placement control design (Controller
Design)
• There are three method:
• Method 1: Direct substitution method (when order of system n≤3)
• Method 2: Bass-Gura Approach
• Method 3: Ackermann’s formula
Controller Design by method 1:
• Let the system is steps are
• Step 1: Check controllability of the system
• Step 2: Put u=-KX where
• So
• Step 3: Write characteristic equations of above new system
• Step 4: Write Desired characteristic equation
• Step 5: Compare above two characteristic equations and solve for k 1, k2,
k3 by equating the power of s on both sides
Controller Design by method 2:
• Let the system is steps are
• Step 1: Check controllability of the system
• Step 2: Put u=-KX where
• Step 3: Let the system is in first companion form (Controllable
canonical form) i.e
Controller Design by method 2…
• Step 4: after putting the value of u in given system, now system will
become . So
---(4)
Controller Design by method 2…
• Step 5:
---(5)
Controller Design by method 2…
• Step 6: Comparing equations (4) & (5) we have
What if the system is not given in first
companion form?
• Answer is to convert it into Companion Form as follows
• Define a transform
• put the value of
• Select the value of T such that is in first companion form
• Put T=MW
• Where is the controllability matrix
What if the system is not given in first
companion form?...
Controller Design using Method 2: Bass-
Gura Approach
• Step 1: Check controllability of the system
• Step 2: Form the characteristic equation using matrix A
• …+ find ai’s
• Step 3: find the transformation matrix T if system is not in first companion T=MW
• Step 4: Write the desired characteristic equation
• Step 5: The required state feedback matrix is
• Note: Above approach is for any system (controllable canonical form or not). If
system is in controllable canonical form put T=I (identity matrix)
Controller Design using Method
3:Ackermann’s Formula
• Let
• Desired characteristic equation
•
• Caley-Hamilton theorem states that every matrix A satisfies it own characteristic
equation. So
• For case n=3 consider the following identities
Controller Design using Method
3:Ackermann’s Formula …
• Multiplying the above identities with respectively and adding them
---(6)
Controller Design using Method
3:Ackermann’s Formula …
• From Caley-Hemilton theorem for
• Also we have for A
• Putting the values & of in equation (6)
0
Controller Design using Method
3:Ackermann’s Formula …
• =>
• Since system is completely controllable inverse of the controllability
matrix exists we obtain
=> ---(7)
Controller Design using Method
3:Ackermann’s Formula …
• Pre multiplying both sides of the equation (2) with [0 0 1]
Controller Design using Method
3:Ackermann’s Formula …
• Hence
• For an arbitrary positive integer n ( number of states) Ackermann’s
formula for the state feedback gain matrix K is given by
are the coefficients of desired characteristic polynomial
Example
• Example 1: Consider the system defined by where
•
• By using the state feedback control u=-KX, it is desired to the closed
loop poles at and s=-10. Determine the sate feedback gain matrix K.
• Solution:
• First check the controllability of above system
Example ..
• Controllability matrix
• so rank of M =3. Hence system is completely state controllable.
• Now we will solve this problem with previous three methods
Example ..
• Method 1: Direct substitution method
• Put u=-KX where
• So
• Write characteristic equations of above new system
Example…
• Desired Characteristic equation
• comparing above two characteristic equations
• k1 = 199, k2 = 55, k3 = 8
• So
Example
• Method 2: Characteristic equation of the given system
• Comparing with
• a1 = 6, a2 = 5, a3 = 1
Example…
• Desired Characteristic equation
• Sate feedback gain matrix K is
•
• Where T= I (identity matrix as system is in controllable canonical form)
Example…
• Method 3: Ackermann’s Formula
Example…
• So
Choice of closed loop poles:
• Don’t choose the closed loop poles far away from the open loop
poles, otherwise it will damage high control effort.
• Don’t choose the closed loop poles very negative, otherwise the
system will be fast reacting (i.e it will have a small time constant)
• In frequency domain it will lead to large bandwidth and hence noise get
amplified.
Controller for multi input system
• The state feedback gain matrix (K) becomes a matrix of mxn (Not
vector of 1xn unlike single input system)
• m = no of inputs and n = no of states
• The state feedback gain matrix (K) is not unique
Summary wise
• Define a linear combination of
control variables as new control
cariable. i.e
• Figure Reference:
• [Link]
Next: Observer Design
Thanks
