HOME  |  NEWS  |  BLOGS  |  MESSAGES  |  FEATURES  |  VIDEOS  |  WEBINARS  |  INDUSTRIES  |  FOCUS ON FUNDAMENTALS
  |  REGISTER  |  LOGIN  |  HELP
Comments
You must login to participate in this chat. Please login.

Good presentation! Thank you Sir

Iron

see you tomorrow again, i am gonna sleep now...too tired from my job. i just scheduled myself to attend the archived class in order to SECURE A PDH CONTINUING EDUCATION CERTIFICATE...

GREAT PRESENTATION Sir Max...much appreciated...thanks to digikey and design news for the events...

much appreciated sir max...great!

2 slides to go then the discussion will be over...nice presentation sir max...

i am now on slide 2...

just downloading the slides presentation for discussion...

good evening again sir max for day 3 lesson!

See you at next session...

Good overview of some algorithms in device control.

People should also look at scheduling systems that "self organize". Holonic systems for example.

Iron

Good presentation again.  Improvement could be made by 'self documenting' the slide presentation with more text rather than verbal descriptions and a few more strategic graphics.

LabView implements most of the mentioned algorithms, some come with tutorials and toolkits, some are white papers, others are programs offered from third parties.  Delight yourself with a "Labview <algorithm>" search on your favorite web browser.  Note: <algorithm> is generic, insert PID, FUZZY, GENETIC and etc

Is there a good book on Fuzzy logic for beginners?

Iron

Presentation was good, thanks

Iron

good presentation, thanks again.

Iron

Electronics today, has better ways to design these systems

Iron

Multiple loops is essential for feed back and control systems

Iron

e-C U Max and everyone tomorrow. Good material and presentation.

Iron

Thanks again Max, have a great time

Iron

A list of references could be compiled by Alex to be e-mailed to all participants.

Iron

I think it is important to be aware that these are all tools /techniques and no one method is panacea...it all depends on the problem domain

Iron

OK Everyone -- I'm going to grab a bite to eat and then ge tback to work -- I hope to C U all tomorrow... Bye till then ... Max

Blogger

@THasham "Hi Max..I think it will be great if you add a slide on reference books at the end of each lecture or at the end of the sessions..Thanks again.. Great lecture"

It's a bit difficult to add anything to the slides that are already up there -- but I think I can work my way around this -- I'll get something that everyone can access by Friday

Blogger

In the desin of complex control systems with high reliability indices on the spec, such as a nuclear energy powered  submarine, safety is integrated across the board on all phases. However, some of the basic design functions stay from time to time with the team of engineers who focus more on the crude functionality of the various processes.

Iron

@snandu13 "Is the logic built in Fuzzy logic only in Software? I donot know how  Fuzzy Logic is built into the controllers."

You can implement any form of logic in either software or hardware -- a software implementation is more versatile with regard to changing it -- and it can be a lot easier to capture your "intent" -- but a hardware implementation will always be much faster (orders of magnitude in many cases) and consume much less power...

Blogger

Hi Max..I think it will be great if you add a slide on reference books at the end of each lecture or at the end of the sessions..Thanks again.. Great lecture

Iron

Max, don't worry, I'll be here tomorrow too. A great way to spend lunch.

Iron

Let me root aroudn my bookshelves and I'll give out a list of books on all sorts of things on Friday

Blogger

@danlafleur "...see you on Friday"

What about tomorrow? :-)

Blogger

@Clive Maxfield- Ok, I think that I am starting to understand this a little bit better.  Add me to the list of people asking you for more book references on this.  Thanks again for another great lecture. 

Bye all.

Iron

Objective importance of safety vs reliability has different relative importance for SIS  vs basic process control system

Iron

@KevinJam: "Saftey system as in the safety Instrumented systems to implement safety functions on the Basic process controll loop is normally handled as an independent network to avoid potentially common cause failure."

That's right -- my job here is done :-)

Blogger

Mr.E, see you on Friday.

Iron

@island_Al: "Slide 17 mentions a calculus based genetic search. Will this be expanded on this week?"

I wasn't planning on doing so ... I just wanted to use this sklide to establish where evolutionary algorithms and GA fit into th egrand scheme of things...

Blogger

@danlafleur: I believe that the Friday class is particularly focused on safety issues so perhaps this will get addressed more fully then as well.

Iron

Saftey system as in the safety Instrumented systems to implement safety functions on the Basic process controll loop is normally handled as an independent network to avoid potentially common cause failure.

Iron

Is the logic built in Fuzzy logic only in Software? I donot know how  Fuzzy Logic is built into the controllers.

Iron

yeah, I know how to see exactly how many points I have for each track, but like I said, I have 22 points for the first track, and I thought I watched everyday. Is there anyway to see the point breakout per lecture instead of per track?

Iron

@clia  "@Clive Maxfield - So, is GA something that would be most apt to use when not looking for immediate responses?"

I would say you've nailed it -- you use GA to quickly find an optimal solution amongst an almost infinite solution space -- once you have that solution you run with it.

But they you have th efact that in a complex system the problem itself (for example the surrounding environment) may change over time ... in which cas eyo umay wish to have an adaptive system -- and the way th esystem adapts may be based on GA techniques...

Blogger

Hi Max..I would also like to get your recommandaiton on Books..

Thanks..

Iron

Slide 17 mentions a calculus based genetic search. Will this be expanded on this week?

 

Gold

Max, I'm interested in safety critical systems. Info on how GA is applied, or even if it's a good idea is something useful for future projects.

Thanks.

Iron

GREAT LECTURE  MAX. THANKS.

my grades are not correct either..  how are they kept track of???

Thanks Max and Alex..

 

Iron

@Max: Thanks Max I appreciate it.

@Mazianlab: "...if you don't mind please give us some refrence book for Fuzzy logic."

Let me think about thsi and get back to you tomorrow

Blogger

@Ranjitha: "... i would like to know how best can one increase the response time of a control system especially when you need to handle safety systems?"

Let me ask my friend Adam about this -- he's based in the UK and is an expert on the safety part of things -- and I'll answer you tomorrow

Blogger

Thanks...Excellent lectures...

@Clive Maxfield - So, is GA something that would be most apt to use when not looking for immediate responses?

Iron

Max, if you don't mind please give us some refrence book for Fuzzy logic.

Iron

That is a really good book..An Introduction to Fuzzy Logic for Priactical Applications by Tanaka


Iron

Thanks Max for expanding my reading list and a few things to let my imagination run wild with.

Iron

For listeners who want info on how certificates are given and the points you earn for each class, go here:

http://www.designnews.com/document.asp?doc_id=236445&ltop=yes

Blogger

Interesting Presentation Thx. Max

Iron

@JM Ashcraft - I missed one of the episodes a few weeks ago and was able to follow up with the archived version.  I did receive the 5-point credit for attendance.

Iron

Books I liked on Fuzzy Logic:

An Introduction to Fuzzy Logic for Priactical Applications by Tanaka

Fuzzy Logic by McNeill and Freiberger

 

@jrjohns: "Given the non-deterministic nature of GA, it seems unlikely that you will see GA used in critical safety systems (or is there a way to get this type of system approved?)"

At the moment I think GA is more of interest in coming up with optimal solutions to a system -- by which I mean more in setting the system up -- as opposed to being used "on th efly" while the system is running

Blogger

@Max: I agree this is a deep topic, but i would like to know how best can one increase the response time of a control system especially when you need to handle safety systems?

Thanks Max.  Another superb seminar.  Very much enjoyed the intro to Genetic Algorithms.  Something I have little experience with.  Will be back tomorrow.

 

jrjohns, it seems to me that safety critical may not be a good GA application as well. try running in every way possible, remember the sucessful one and do it again under different conditions.

Iron

is there any way to see which lessons we have credit for? because i watched all 5 of the first track and i only have 22 points

 

Iron

archives - can you get the extra 5 points?

There are alot of books on algorithums and these higher level algorithums lead you toward these types of calculations

Great session again today - thanks Max!

Iron

Thank you for today's lecture

Iron

Is the logic built in Fuzzy logic only in Software? I donot know how is Fuzzy Logic is built in controllers.

Iron

Thanks Max, Alex, Clive, and Frank.

Iron

Given the non-deterministic nature of GA, it seems unlikely that you will see GA used in critical safety systems (or is there a way to get this type of system approved?)

Iron

bye all, till tomorrow

Iron

thanks Max for a good presentation

Iron

"primitive" solutions with straightforward programs are comparable to the amoebas ... telling the spider to "figure out how to walk" is quite another level

Iron

Not only are their cookbooks, but how about books with kits and software for learning?

Iron

@caa028- Is the mathematical proof significant?  I don't mean to split hairs, but are we talking about it finding the solution after only 3 million attempts as opposed to 3 million and 5 attempts?

Iron

Max, when you implement GAs, do you start with a toolkit or starting point of some sort?  Is it free or for sale?

Iron

Max, interesting GA robot walking case.  Have you considered using the same approach for a bipedal robot walking and/or standing up?

Iron

JM: He referred to it as the "F" word.

Iron

Thanks for a great presentation Max.

What about FPGAs, never mentioned once

You did not mention calculus solutions?

Gold

Thanks Max, a bit confusing but interesting non the less

Iron

clia: if you used millions of blind searchers, the chance of finding what you're looking for, are growing

Iron

@clia - matematically proven to be more efficient that random search, sequential search, etc...

Iron

I've used dynamic fuzzy logic for robotic control..

Iron

The problem with GAs is making a good testing algorithm.  It becomes an escalation between you the code writer and your GA

Iron

@Kanoto - What is it in the Motorola HCS12 uC that orients it toward fuzzy logic?

Iron

The tool/technique used depend on your problem space requirements

Iron

Max: What did you say the lab in Oslo is called?

Iron

Fascinating yes, but it seems just a hair less hit and miss than blindly searching, no?

 

Iron

I would hate to debug this

GA sounds like the first attempts at troubleshooting a system. Try random things, remember what workd and try variations of that until you have a set of recognized patterns and solutions.

Iron

GAs also are often very fault tolerant with their solutions

Iron

Genetic algoritms..be careful of local minima /maxima not optimal solution

Iron

Testing 1...  2...  3...

Iron

clia: the merit is that "evolution" might find answers you would not have thought of using a "logical" approach

 

Iron

It is a fascinating concept.

Platinum

GAs let you search a VERY large space efficiently for solutions

Iron

@KevinJam & Greeng - if I did not know anything about AI, I might have believed that you were serious... garbage in, garbage out...

Iron

Very interesting these Genentic Algorithm methodology for problem solving.

Iron

Is anyone else not exactly clear on the merit of Genetic Algorithms?

Iron

As a rule employ simplest least cost method that you can...don't add more complexity to solution

Iron

Artificial life by Steven Levy is a very enjoyable book on Genetic Programming/Alogorithms

I am not seeing any GA discussion.  Does anyone out there use it or know anything about it?

Iron

As per Fuzzy. most errors occur around "fussification..

Iron

Fuzzy can be computationally intensive as variable inputs increase

Iron

@kenstan -- neural network is essentially a network of MACs with thresholds at the outputs

Iron

cost increases with complexity and level of computation required, along with comensurate hardware

Iron

Sorry is was too "

what ia neural network?

Iron
Iron

using simple Min-max inference engines

Iron

I use Fuzzy for robotic gripper control when interfacing with humans

Iron

Fuzzy is the way to go....

Iron

sorry increasing response times..

Iron

Oh no, thank god it's limited to 5 points..

Iron

@Greekeng - a perfect answer to a "what" question...

Iron

@Greekeng- Ha Ha!

 

Iron

Are there any tools or libraries for GA development in HW?

Iron

@kenstan, they are used on the NCC-74656

Iron

like i said it depends on the nature of yourproblem

Iron

how about using a divide and conquer approach to map the solution space? I guess you have to have time to do this, oh, and a non-destructing system

Iron

Slide 16 represents a comon problem in engineering optimization problems, such as optimal control design of sychronous machines, which are highly non-linear.

Iron

the best approach is to understand the nature of the type of process being controlled / its inherent behaviour before choosing a control strategy...it is possible you may well compose a combination of stragy approaches

Iron

@KevinJam I was able to compress video using bounded 4 steps genetic algo with very good result.

 

Iron

what ia neural network?

Iron

@KevinJam- So, the more you know about it, the easier it is to teach/learn?

Iron

Controls require time bounded response

Iron

Motorola HCS12 uC have fuzzy logic implementation in them.

Iron

Combination of learning algorithms based on nature of data if known

 

Iron

@KevinJam - in real life learning also takes time...

Iron

@KevinJam- Good point.  And, the more that you want it to know the longer it is going to take.  Has any work been done to speed this process up?

Iron

i recall articles years ago describing something like an FPGA that changed the circuit in a dynamic way according to the signal applied. signal characteristic or header info.

Iron

Neural networks have been widely used in pattern recognition algorithms such as OCR.

Iron

Neural networks & FPGAs....too hard, too limited, too slow, and takes a lot of gates and power

Iron

with neural networks learning takes time

Iron

WEKA (http://www.cs.waikato.ac.nz/ml/weka/) is an interesting tool to play with to learn about AI 

Iron

Combine fuzzy logic and neural networks - adapt the membeship functions to match input semantics

Iron

has anybody done anything useful with a neural network in a FPGA?

Iron

no Indiana Jones ref yet

Iron

No, Ward's just trying to create one

Iron

Was there an Indiana reference today??  I missed the first few minutes

Iron

Wasn't Indiana Jones inspired by Monty Python?

Iron

it appears that most electronics in modeling real world organisms???

Max, thanks. Wetware sounds like really interesting reading.

Iron

A computer in every living cell

Iron

Wet Ware... (subtitle was?)

Iron

Neural networks - supervised and unsupervised learning

Iron

thanks everyone for recommending books

Iron

thanks, skechy

found it too ... buried below hair product ... time to find a real bookstore, I guess

Iron

The integral, due to its tendency to facilitate lag..affords overshoot

Iron

@caa028- Ahh, thank you.

Iron

@clia - I don't think, he was talking about control or fuzzy logic... ;)

Iron

Derivative increases speed of response

Iron

PID: P is based on now, I is based on the recent past, D is based on near future.

Iron

PID doesnt always eliminate the overshoot... it helps to tune if you want faster responce or no overshoot or something inbetween

Iron

@caa028 - What is the difference?

Iron

Diazo & Houpis or Dorf have good books on Controls

Iron

using simple Min-max inference engines

Iron

It is important to note that you should have good understanding of process to be controlled before application of any of these techniques

Iron

@KevinJam, Doesn't the addition of the derivative (slope of the error) lead to the possibility of overshoot?

Iron

@Greekeng - "fuzzy" or "fuzzy logic"?

Iron

I used Morris Driels for control systems.. great book.. but Nice, Ogata, Dorf are other popular ones

 

Iron

Max: Actually, the notion of model based design is not only inherant to Model Predictive Control, but has been with P, PD, PID since before the II world War in practical applications such as syc machine control and rocket launching and navigational control.

Iron

any book do you recommend about adaptive control?

Iron

as a friend of mine use to say; '...for all the programs i've writen, i was always using {their} 'fuzzy-logic'.

schmidl: Art of control

Iron

I use Fuzzy for robotic gripper control when interfacing with humans

Iron

closed loop and adaptive control are little different..adaptive has more loops

 

Iron

for oscillation...synonymous with ringing

Iron

PID eliminates the overshot.

Iron

does anybody have some good book recommendation on general control system theory?

Iron

in general, adaptive controller has two closed loops.

Iron

1. There are specific modelling and control algorithms (techniques) for event driven systems also known as discrete event systems. Petri nets is a well known approach for these kind of systems.

2. Could you recommend a Petri Net simulator (or tool) that is able to generate code for certain processors (microcontrollers, FPGA, etc.)? Thank you.

 

Iron

@KevinJam, you mean it won't ring/overshoot?

Iron

All being discussed are closed loop

Iron

look into Zeigler nicholas tuning for pid

Iron

@EugeneM - Just one example of a closed loop, but yes

 

Iron

PID inherently liniear controller

Iron

adaptive control = closed loop?

Iron

Are there any basic hints for getting the PID control right quickly so that it is fast without ringing?

Iron

you can also consider that the event is something dinamic and state something static

Iron

first order modelled response -Desireable ; fast, minimal oscillation and minimal lag 

Iron

@MaxianLab: an event occurs at a point in time and then is over, but a state is a condition of the system which can change the interpretation of incoming events. For example, a switch closure would be an event and might or might not cause a state change. Initialization might be a state in which such switch closures are ignored, while normal operation might be a state where the switch closures are detected and responded to somehow (perhaps with a change to yet another state).

Iron

@Maian, an event is triggered by something (time, sensor reading, switch, etc) wereas state is internal to the process keeping track of progress and what happens next.

Iron

@MazianLab: Events are the things that cause transitions between states.

Iron

@MazianLab: an event happens at a moment in time, a state is what lies between these

Iron

@jrjohns - Got it.  Thanks for example.

Iron

@caa028, Yeah, but your up to two points already and catching up nicely.  :)

Iron

what is the diference between event and state?

Iron

Clia - Think thermostat.  It is either on or off to control temperature

Iron

oh, i forgot to say "Hi There"

Iron

Just a few minutes late today and it sounds like I've already missed all the fun...

Iron

Yeah, I am still a little fuzzy on Bang-Bang control

Iron

also modulating control

Iron

sorry increasing response times..

Iron

I think that "Bang-Bang" control was used on the Mars lander? ;-)

Iron

I have not tested. Go ahead...

@slk - Really?  Counts for me.

Iron

One big issue when switching from Analog to digital, the propagation delays that are added, thus reducing response times, big issue in drive systems.

Iron

You can also comment after the days lesson

@Bruce, have you tested this? 

Iron

hi  - doesn't count

Iron

one point per post, max 5

Iron

you can check by clicking "my grades" link on top

Iron

@jl - I believe you would need 5

 

Iron

how many "hi there's" do you have to post before you get the 5 extra points.

Iron

You haven't really programmed a computer until you have set at a console with 24 switches to set the 0s and 1s for a word and pushed another switch to enter each word in memory before running the program.

Iron

Oh no, thank god it's limited to 5 points..

Iron

@danlafleur: Actually computers are like ties - old ideas (think virtual machines) come back in style!

 

Iron

now we will get lots of "Hi there"

Iron

Although I think the idea was the chat was actually supposed to be about the lecture...

Iron

now that is what I am talking about!

no extra credit past a certain point, just fun.

Iron

Actually yes, up to 5 points per session :-)

Iron

@JM Ashcraft. yes you do ;)

 

Iron

I remember that with a whole lot of bnc connectors

do you get extra credit for this chit chat?

Hi everyone, late but glad to be here!

Iron

IBM 360 Model 44 - Still have the green card at home :-)

Iron

wow, we're in good company with dark age computing.

Iron

IBM360 Assembly here :-)

Iron

don't know why we are spending so much time repeating what was presented yesterday.

Iron

Some subsystems can be linearized

Iron

Fortran & punch cards here too.

Iron

Feedforward - form of open loop

Iron

densgt, it's good to remember your roots. learn it the hard way and remember forever.

Iron

@ danlafleur- I'm showing my age, but my first (Fortran) course used punch cards.

Iron

ASR33 and paper tape was my first computer math couse experience.

Iron

Alex, did you use 5 bit code as well

Iron

@ danlafleur - that "tune" is #1 in the top.  can do it all.

Iron

good afternoon everyone

Iron

good day all. good to be back

Iron

could never compete with Max, it's just like having a tune stuck in your head :-)

Iron

Good afternoon. Just cught up on pts 1 & 2 after taking a long weekend.

Iron

Type of control strategy employed would depend on the problem

Iron

Dan, are you competing with Max on the number of times you'll use FPGA today?

Iron

Max, do you find that most of the controls work you do today is so called "modern" (state variable) control or do you find yourself doing mostly classical control?

Iron

Possible analogy is to suggest DCS is Process Oriented or time driven, whereas SCADA is Data Oriented or event driven. Actual control laws could utilize any appropriate method - time or event

Iron

Good morning and a fine FPGA day to all.

Iron

Students:  Note your "experience", plz

Please join our Digi-Key Continuing Education Center LinkedIn Group at http://linkd.in/yoNGeY

Blogger

Be sure to click 'Today's Slide Deck' under Special Educational Materials above right to download the PowerPoint for today's session.

Blogger

The streaming audio player will appear on this web page when the show starts at 2pm eastern today. Note however that some companies block live audio streams. If when the show starts you don't hear any audio, try refreshing your browser.

Blogger


Partner Zone
Latest Analysis
This Gadget Freak review looks at a cooler that is essentially a party on wheels with a built-in blender, Bluetooth speaker, and USB charger. We also look at a sustainable, rotating wireless smartphone charger.
Texas Instruments is rolling out a new microcontroller that could make the design of sensor networks and data logging systems simpler and less costly.
Made By Monkeys highlights products that somehow slipped by the QC cops.
From pitchers and forwards to quarterbacks and defensemen, we offer a peek at some of the more memorable engineers in sports history.
IBM announced it is dedicating $3 billion of funding over the next five years to research and development of new processor technologies.
More:Blogs|News
Design News Webinar Series
7/23/2014 11:00 a.m. California / 2:00 p.m. New York
7/17/2014 11:00 a.m. California / 2:00 p.m. New York
6/25/2014 11:00 a.m. California / 2:00 p.m. New York
5/13/2014 10:00 a.m. California / 1:00 p.m. New York / 6:00 p.m. London
Quick Poll
The Continuing Education Center offers engineers an entirely new way to get the education they need to formulate next-generation solutions.
Jul 21 - 25, Design Products With Bluetooth Low Energy
SEMESTERS: 1  |  2  |  3  |  4  |  5  |  6


Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.
Next Class: August 12 - 14
Sponsored by igus
Learn More   |   Login   |   Archived Classes
Twitter Feed
Design News Twitter Feed
Like Us on Facebook

Sponsored Content

Technology Marketplace

Copyright © 2014 UBM Canon, A UBM company, All rights reserved. Privacy Policy | Terms of Service