Department of CONTENTS

Automation and Robotics

 Proportional (P), Integral (I) and Derivative (D) control
actions
 PI, PD and PID control systems in parallel form
 Unit step Response analysis via Transient response
UNIT NO.6: specifications: Percentage overshoot, Rise time, Delay
CONTROL SYSTEMS time, Steady state error
 Manual tuning of PID control
 Linear Quadratic Control (LQR)

Presented by: Prof. D. P. Patil

INTRODUCTION  With the use of a low cost simple ON-OFF controller, only
two control states are possible, like fully ON or fully OFF. It
PID controllers are found in a wide range of applications for is used for a limited control application where these two
industrial process control. Approximately 95% of the closed- control states are enough for the control objective.
loop operations of the industrial automation sector use PID However oscillating nature of this control limits its usage
controllers. and hence it is being replaced by PID controllers.
PID stands for Proportional-Integral-Derivative. These three  PID controller maintains the output such that there is zero
controllers are combined in such a way that it produces a error between the process variable and setpoint/ desired
control signal. As a feedback controller, it delivers the output by closed-loop operations. PID uses three basic
control output at desired levels. control behaviors that are explained below.
Before microprocessors were invented, PID control was
 A PID controller continuously calculates an error value
implemented by the analog electronic components. But
today all PID controllers are processed by the e(t) as the difference between a desired setpoint (SP)
microprocessors. and a measured process variable (PV) and applies a
correction based on proportional, integral,
Programmable logic controllers also have the inbuilt PID and derivative terms (denoted P, I, and D respectively),
controller instructions. Due to the flexibility and reliability of hence the name.
the PID controllers, these are traditionally used in process
control applications.

1
PROPORTIONAL (P) CONTROLLER  This controller requires biasing or manual reset
when used alone. This is because it never reaches
 Proportional or P- controller gives an output that is the steady-state condition.
proportional to current error e (t).
 It provides stable operation but always maintains
 It compares the desired or set point with the actual value the steady-state error.
or feedback process value.
 The speed of the response is increased when the
 The resulting error is multiplied with a proportional
constant to get the output.
proportional constant Kc increases.
 If the error value is zero, then this controller output is zero.

PROPORTIONAL CONTROLLERS PROPORTIONAL CONTROLLERS

 Now we are in a condition to discuss proportional
controllers, as the name suggests in a proportional  Where Kp is proportional constant also known as
controller the output (also called the actuating controller gain.
signal) is directly proportional to the error signal.  It is recommended that Kp should be kept greater
 Now let us analyze the proportional controller than unity.
mathematically. As we know in proportional  If the value of Kp is greater than unity (>1), then it
controller output is directly proportional to the error will amplify the error signal and thus the amplified
signal, writing this mathematically we have, error signal can be detected easily.

• Removing the sign of proportionality we have,

2
ADVANTAGES OF PROPORTIONAL CONTROLLER DISADVANTAGES OF PROPORTIONAL CONTROLLER
 The proportional controller helps in reducing the  Due to the presence of these controllers, we get
steady-state error, thus makes the system more some offsets in the system.
stable.
 Proportional controllers also increase the maximum
 The slow response of the overdamped system can be overshoot of the system.
made faster with the help of these controllers.

INTEGRAL CONTROLLERS
INTEGRAL CONTROLLERS
 Due to the limitation of p-controller where there  As the name suggests in integral controllers the
always exists an offset between the process variable output (also called the actuating signal) is directly
and setpoint, I-controller is needed, which provides proportional to the integral of the error signal. Now
necessary action to eliminate the steady-state error. let us analyze integral controller mathematically.
 It integrates the error over a period of time until the  As we know in an integral controller output is
error value reaches zero. It holds the value to the directly proportional to the integration of the error
final control device at which error becomes zero. signal, writing this mathematically we have,

 Removing the sign of proportionality we have,

3
 Where Ki is an integral constant also known as ADVANTAGES OF INTEGRAL CONTROLLER
controller gain. The integral controller is also
known as reset controller.  Due to their unique ability, Integral Controllers can
return the controlled variable back to the exact set
 Integral control decreases its output when a point following a disturbance that’s why these are
negative error takes place. It limits the speed of known as reset controllers.
response and affects the stability of the system. The
speed of the response is increased by decreasing
integral gain, Ki.

DISADVANTAGES OF INTEGRAL CONTROLLER DERIVATIVE CONTROLLERS

 We never use derivative controllers alone. It should
 It tends to make the system unstable because it be used in combinations with other modes of
responds slowly towards the produced error. controllers because of its few disadvantages which
are written below:
 It never improves the steady-state error.
 It produces saturation effects and also amplifies the
noise signals produced in the system.
 Now, as the name suggests in a derivative controller
the output (also called the actuating signal) is
directly proportional to the derivative of the error
signal.

4
DERIVATIVE CONTROLLERS
ADVANTAGES OF DERIVATIVE CONTROLLER
 Now let us analyze the derivative controller
mathematically. As we know in a derivative  The major advantage of a derivative controller is that
controller output is directly proportional to the it improves the transient response of the system.
derivative of the error signal, writing this
mathematically we have,

 Removing the sign of proportionality we have,

 Where, Kd is proportional constant also known as

controller gain. The derivative controller is also
known as the rate controller.

 Removing the sign of proportionality we have,

PROPORTIONAL AND INTEGRAL CONTROLLER
 As the name suggests it is a combination of
proportional and an integral controller the output
(also called the actuating signal) is equal to the  Where, Ki and kp are proportional constant and integral
summation of proportional and integral of the error constant respectively.
signal.  Advantages and disadvantages are combinations of the
 Now let us analyze proportional and integral advantages and disadvantages of proportional and
controller mathematically. integral controllers.
 Through the PI controller, we are adding one pole at
 As we know in a proportional and integral controller
origin and one zero somewhere away from the origin (in
output is directly proportional to the summation of the left-hand side of complex plane).
proportional of error and integration of the error
signal, writing this mathematically we have,  As the pole is at the origin, its effect will be more, hence
PI controller may reduce the stability; but its main
advantage is that it reduces steady-state error drastically,
due for this reason it is one of the most widely used
controllers.

5
PROPORTIONAL AND INTEGRAL CONTROLLER
 The schematic diagram of the PI controller Against step
input, For the values of K=5.8, Ki=0.2, Its time response, At
K=5.8 (As a P- controller, it was on the verge of instability,
so just by adding the small value of an Integral part, it
became unstable.
 Please note the Integral part reduces the stability, which
does not mean that system will be always unstable. In the
present case, we have added an integral part and the
system became unstable).
The response of the system shown in above figure with K=5.8, Ki=0.2

The closed loop control system with PI Controller

PROPORTIONAL AND DERIVATIVE CONTROLLER  Where, Kd and Kp are proportional constant and
derivative constant respectively.
 As the name suggests it is a combination of
proportional and a derivative controller the output  Advantages and disadvantages are combinations of
(also called the actuating signal) is equals to the advantages and disadvantages of proportional and
summation of proportional and derivative of the error derivative controllers.
signal. Now let us analyze proportional and derivative  Readers should note that adding ‘zero’ at the proper
controller mathematically. location in the open-loop transfer function improves
 As we know in a proportional and derivative controller stability, while the addition of pole in the open-loop
output is directly proportional to the summation of transfer function may reduce the stability.
proportional of error and differentiation of the error  The words “at proper location” in the above sentence
signal, writing this mathematically we have, are very important & it is called designing of the
control system (i.e. both zero & pole should be added
at proper points in the complex plane to get the
desired result).
 Removing the sign of proportionality we have,  Inserting the PD controller is like the addition of zero in
open-loop transfer function [G(s)H(s)]. Diagram of PD
Controller is shown in below figure.

6
 Closed-loop control system with PD Controller

Response of system shown in above figure with K=5.8, Td=0.5

 In the present case, we have taken the values of K=5.8,
Td=0.5. Its time response, against step input, is shown in
below figure

 The transfer function of the PD controller is K+Tds or PROPORTIONAL PLUS INTEGRAL PLUS DERIVATIVE
Td(s+K/Td); so we have added one zero at -K/Td. By CONTROLLER (PID CONTROLLER)
controlling the value of ‘K’ or ‘Td’, the position of the
‘zero’ can be decided.
 A PID controller is generally used in industrial control
 If ‘zero’ is very far away from the imaginary axis, its
influence will decrease, if ‘zero’ is on the imaginary applications to regulate temperature, flow, pressure,
axis (or very close to the imaginary axis) it will also be speed, and other process variables.
not accepted (root locus generally starts from ‘poles’ &  A type of controller in which the output of the
terminates at ‘zero’, Designer’s aim is generally such controller varies in proportion with the error signal,
that root locus should not go towards the imaginary integral of the error signal and derivative of the error
axis, due to this reason ‘zero’ very near to imaginary signal is known as the proportional integral derivative
axis is also not acceptable, hence a moderate position controller. PID is the acronym used for this type of
of ‘zero’ should be kept) controller.
 Generally,it is said, PD controller improves transient  Proportional plus integral plus derivative controller is
performance and the PI controller improves the steady-state sometimes referred as a 3-mode controller, as it
performance of a control system.
combines the controlling action of proportional,
integral as well as derivative controller altogether.

7
  The combination of all the three types of control action
improves the overall performance of the control system,
in order to provide the desired output in an effective
manner. The output of a PID controller is given as:

 Here we have a block diagram of the PID controller:

 Here Kp is proportionality constant for the error signal,  We know that the transfer function is represented as
 Ki is the proportionality constant for integral of the output by input. Thus the transfer of the controller will
error signal and be given as,
 Kd is the proportionality constant for the derivative of
the error signal.
 Further to determine the transfer function of the  On further simplification
controller, the time domain function must be converted
to the frequency domain. Therefore, considering the
Laplace transform of the above equation, we will get

 Consider, Kp/Ki = Ti and Kd/Kp = Td and substitute in

the equation given above, we will get,

8
  So, further

 Since, Kp/Ki = Ti therefore Kp/Ti = Ki

 And Kd/Kp = Td, therefore, Kd = Kp Td

Closed loop control system with PID Controller

 The transfer function of the PID Controller can be

 This is given as the simplified transfer function of found as:
the PID controller.

 It can be observed that one pole at origin is fixed,

remaining parameters Td, K, and Ki decide the position
of two zeros.
 In this case, we can keep two complex zeros or two real
zeros as per the requirement, hence PID controller can
provide better tuning. In the olden days, the PI
controller was one of the best choice of control
engineers, because designing (tuning of parameters)
of the PID controller was a little difficult, but nowadays,
Response of system shown in above figure, with K=5.8, Td=0.5, Ki=0.2
due to the development of software designing of PID
controllers have become an easy task.
 Against step input, For the values of K=5.8, Ki=0.2, and
Td=0.5, Its time response is shown in below figure.

9
 Response of PV to step change of SP vs time, for Response of PV to step change of SP vs time, for
three values of Kp (Ki and Kd held constant) three values of Ki (Kp and Kd held constant)

UNIT STEP RESPONSE ANALYSIS VIA TRANSIENT RESPONSE

SPECIFICATIONS: PERCENTAGE OVERSHOOT, RISE TIME,
DELAY TIME, STEADY STATE ERROR

Response of PV to step change of SP vs time, for

three values of Kd (Kp and Ki held constant)

10
MANUAL TUNING OF PID CONTROL Effects of increasing a parameter independently

 The determination of corresponding PID parameter values for

getting the optimum performance from the process is called
tuning.

11
LINEAR QUADRATIC CONTROL (LQR)  The cost function is often defined as a sum of the
deviations of key measurements, like altitude or
 The theory of optimal control is concerned with process temperature, from their desired values. The
operating a dynamic system at minimum cost. The case algorithm thus finds those controller settings that
where the system dynamics are described by a set minimize undesired deviations. The magnitude of the
of linear differential equations and the cost is control action itself may also be included in the cost
described by a quadratic function is called the LQ function.
problem. One of the main results in the theory is that
the solution is provided by the linear–quadratic  The LQR algorithm reduces the amount of work done
regulator (LQR), a feedback controller whose by the control systems engineer to optimize the
equations are given below. controller. However, the engineer still needs to specify
the cost function parameters, and compare the results
 The LQR can be run repeatedly with a receding with the specified design goals. Often this means that
horizon; this is a form of model predictive control. controller construction will be an iterative process in
 The settings of a (regulating) controller governing which the engineer judges the "optimal" controllers
either a machine or process (like an airplane or produced through simulation and then adjusts the
chemical reactor) are found by using a mathematical parameters to produce a controller more consistent
algorithm that minimizes a cost function with weighting with design goals.
factors supplied by a human (engineer).

 While solving the dynamic programming problem for

continuous systems is very hard in general, there are a
few very important special cases where the solutions
are very accessible. Most of these involve variants on
the case of linear dynamics and quadratic cost. The
simplest case, called the linear quadratic regulator
(LQR), is formulated as stabilizing a time-invariant
linear system to the origin.
 The linear quadratic regulator is likely the most
important and influential result in optimal control
theory to date. In this chapter we will derive the basic
algorithm and a variety of useful extensions.
 The Linear Quadratic Regulator (LQR) is a well-known
method that provides optimally controlled feedback
gains to enable the closed-loop stable and high
performance design of systems.

12
13
 In a real-world application, it is common for a robot to use
the Extended Kalman Filter to calculate near-optimal
estimates of the state of a robotic system and to use LQR to
generate the control values that move the robot from one
state to the next

 Control commands are sent by the LQR algorithm

 The robot changes state
 Sensor measurements are made
 The sensor measurements are used to generate near-
optimal estimates of the state
 These estimates of the state get fed into the LQR
controller to determine what the next control commands
should be

14