Predictive control 1 introduction

Predictive control 1
Introduction
Anthony Rossiter
1
Slides by Anthony Rossiter

What is predictive control?
Predictive control technique is very widely
implemented within industry and hence of large
interest to all graduates for whom control may
form part of their duties.
Key aspects that students need to appreciate are
that:
1. Predictive control describes an ‘approach’ to
control design, not a specific algorithm.
2. A user would ideally interpret the approach to
define an algorithm suitable for their own
needs.
2
An off-the shelf algorithm may not
be a good choice in many cases!

Implications
It is less important that students learn specific
details of a given predictive control algorithm.
It is more important that they understand key
concepts.
1. Why is it done this way?
2. What is the impact of uncertainty?
3. How does this change with constraints?
4. What are the ‘tuning’ parameters and how can
I use them? Which choices are poor and why?
5. Etc.
3

This lecture series
The main focus is on concepts.
• Why is predictive control logical and systematic?
• What are the consequences of taking a
systematic control approach or concept and
turning into the mathematics required for a
computer implementation.
• What assumptions and design choices are
implicit in a well thought through algorithm?
4
The key to effective implementation is a good
understanding of how MPC works!

Why is predictive control logical?
Many effective control strategies have their origins
in human behaviour. Moreover, humans are very
good at control so a good start point for
automation techniques!
• PID can be deconstructed as a simplification of a
human technique for controlling simple systems.
• Similarly, the use of predictions of expected
behaviour in determining a control strategy is
intuitively obvious.
Some examples will make this clear.
5

Driving a car
6
What is a core component of driving?
1. Drivers look ahead and anticipate future ‘targets’
or demands.
2. Change in the road, pedestrians, other vehicles,
change in speed limit, etc.
We use anticipation, that is prediction, to help
determine effective control strategies.

Racquet sports
7
What is a core component of badminton?
Watch the video!
1. Players plan several shots ahead in order to move
their opponent into a weaker position, or to
prevent themselves being put in such a position.
2. They predict the impact of different shot choices,
and select the ones which lead to the most
desirable outcome.

Filling a tank to a desired level
8
What is a core component of human strategies for
level control?
1. We observe the change in depth and anticipate the
future changes.
2. We modify the input flow to ensure the future
depth does not exceed the target.

Tank fill example
Let us consider in more detail what the human is
doing in the tank example.
9
0 1 2 3 4 5
0
0.5
1
1.5
2
2.5
time(sec)
depth
Predicted evolution
of depth at t=3 for
constant flow.
What action will prevent
over filling the tank?
Update your predictions.

Interim summary
• It is clear that humans use anticipation,
effectively prediction, in order to consider the
impacts of different control strategies.
• They choose the strategy they expect to give the
most desirable future outcome.
PREDICTION UNDERPINS PRACTICAL HUMAN
CONTROL STRATEGIES AND THUS SEEMS A
LOGICAL CONCEPT TO INCORPORATE INTO
AUTOMATED STRATEGIES.
10

Why predictive control?
This video has given a superficial view on why the use
of prediction seems logical, but one can list many
other advantages of predictive control which motivates
further study. Some main ones are:
1. Intuitive concept, easy to understand and
implement for a variety of systems.
2. Systematic handling of constraints.
3. Handles MIMO systems and dead-time without any
modification.
4. Feed forward to make good use of future target
information is included implicitly.
5. Handles challenging dynamics (unlike PID).
11

Why popular in industry?
A simple answer is that it has been proven to
improve profits by giving superior control
compared to conventional techniques.
1. A typical argument is that, if one is confident
that the variance of the output can be reduced,
one can then safely operate closer to a
constraint and therefore increase output
quantity or quality.
2. The ability to incorporate constraints explicitly
enables ‘optimum’ constrained performance as
opposed to the consequences of ad hoc fixes.
12

© 2014 University of Sheffield
This work is licensed under the Creative Commons Attribution 2.0 UK: England & Wales Licence. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/2.0/uk/ or send a letter to: Creative Commons, 171 Second Street, Suite 300, San Francisco, California 94105, USA.
It should be noted that some of the materials contained within this resource are subject to third party rights and any copyright notices must remain with these materials
in the event of reuse or repurposing.
If there are third party images within the resource please do not remove or alter any of the copyright notices or website details shown below the image.
(Please list details of the third party rights contained within this work.
If you include your institutions logo on the cover please include reference to the fact that it is a trade mark and all copyright in that image is reserved.)
Anthony Rossiter
Department of Automatic Control and
Systems Engineering
University of Sheffield
www.shef.ac.uk/acse

Predictive control 1 introduction

More Related Content

What's hot (19)

Similar to Predictive control 1 introduction (20)

Recently uploaded (20)

Predictive control 1 introduction