![funskits](/img/default-banner.jpg)
- Видео 60
- Просмотров 215 999
funskits
Добавлен 8 дек 2006
Shortest-path Grocery Shopping
What's the shortest path through my grocery store that gets all the items on my shopping list? We apply Mathematica's powerful image-processing and graph algorithms to a real map of my local grocery store to figure it out.
More info:
justinppearson.com/projects.html#shortest-path-grocery-shopping
(c) 2020 Justin P. Pearson. All rights reserved.
More info:
justinppearson.com/projects.html#shortest-path-grocery-shopping
(c) 2020 Justin P. Pearson. All rights reserved.
Просмотров: 1 033
Видео
An architecture for preemption-resistant control (Ph.D. Defense)
Просмотров 3476 лет назад
We present a control architecture that allows a controller running on a non-real-time operating system to run with a high degree of determinacy, even in the presence of OS preemption. The main idea is to buffer the sensor measurements and actuator commands on an auxiliary microprocessor called a Real-Time Unit. The RTU runs independently from the main CPU and so can function even if the main co...
Control with minimum energy-per-symbol, and the optimality of event-based control (Ph.D. Defense)
Просмотров 3426 лет назад
The performance of a control system is often limited by constraints on timing, bandwidth, and energy. This work explores the trade-offs between constraints on these resources, the control system performance, and the system to be controlled. We begin by considering a networked control system in which the sensor sends its measurements to the controller over a limited-bandwidth communications chan...
A program to play the Set card game
Просмотров 3816 лет назад
"Set" is a pattern-matching card game. In this project we describe a Mathematica program to play Set over a USB video camera. This requires some interesting image processing, including developing a shape classifier using on a probabilistic machine learning algorithm called multinomial logistic regression. Justin Pearson Thurs Jan 11, 2018 Tech Talk Appfolio, Inc For the code and slides, see: ju...
Load-testing a Rails app with Tsung on AWS Elastic Beanstalk
Просмотров 2,6 тыс.6 лет назад
We deploy a Ruby-on-Rails app to Amazon Web Services "Elastic Beanstalk" and load-test it with Tsung. Table of Contents: 1:02 Demo of Rails app 2:18 Architecture of Rails app 3:57 Deploy to Elastic Beanstalk 9:01 Load-test with Tsung This was a class project for CS 291 Scalable Internet Services at the University of California at Santa Barbara (cs291.com). The class teaches techniques for build...
Teaching University Discussion Sections
Просмотров 4216 лет назад
This workshop helps teachers plan their lessons, engage their students, and assess their teaching. I presented this to new Teaching Assistants (1st-year graduate students) at UC Santa Barbara’s 2017 Teaching Assistant Orientation. Workshop synopsis: Engineering and its practitioners have the reputation for being boring and impenetrable. In this workshop we will discuss ways to structure your le...
Tour of the Controls Lab at UC Santa Barbara
Просмотров 1,9 тыс.7 лет назад
A tour of the Controls Lab at the University of California at Santa Barbara, for incoming first-year undergraduate students in Electrical and Computer Engineering. 0:35 - what is control theory? example: cruise controller 7:40 - inverted pendulum hardware 12:15 - humans try to balance 15:48 - computer balances pendulum 18:00 - balancing with additional setpoints 19:03 - unmodeled vibrational mo...
Zeal 2 [HD]
Просмотров 1077 лет назад
A high-res (1600-by-1200) version of "Zeal 2" our 2004 storyboard sequel to our chase scene "Zeal". "Zeal 2" original low-res: ruclips.net/video/dQPrZc8XIQ8/видео.html "Zeal 1" chase scene: ruclips.net/video/Fgq9dGgKsuE/видео.html Music: "Suite from X2", soundtrack from the motion picture "X-men 2", John Ottman More videos, projects, code, presentations, contact info: justinppearson.com (c) 200...
MapReduce / Paxos class project
Просмотров 4947 лет назад
This video describes my Distributed Systems class project. The project involved implementing the MapReduce parallel computing algorithm and the Multi-Paxos distributed consensus algorithm. The code uses MapReduce to parallelize the computation of a word tally of several large text files; then it uses Multi-Paxos to distribute those tallies among two other computers in the form of a replicated l...
Minimum-Energy Encoding for Networked Control Systems
Просмотров 3767 лет назад
Title: Minimum-Energy Encoding for Networked Control Systems Justin Pearson Oct 25, 2013 25th Southern California Control Workshop Hosted at the University of California, Irvine Abstract: We consider the problem of stabilizing a linear system when communication constraints limit how often the controller learns the state of the system. We introduce a communications framework that accounts for th...
Dustin the Bad TA makes classic teaching mistakes
Просмотров 3,1 тыс.7 лет назад
"Dustin the bad teaching assistant" unwittingly demonstrates classic teaching mistakes during my TA workshop “TAing Engineering Sections”. Sep 22, 2015 Teaching Assistant Orientation University of California, Santa Barbara Workshop synopsis: Engineering and its practitioners have the reputation for being boring and impenetrable. In this workshop we will discuss ways to structure your lessons an...
Paxos in Pictures: Lamport's Distributed Consensus Algorithm
Просмотров 17 тыс.7 лет назад
Paxos is a computer algorithm to help a network of computers agree on a proposed value. It is hard to understand at first. This video attempts to explain Paxos intuitively yet rigorously. The content comes from Lamport's 2001 paper "Paxos Made Simple". Note that this video is not as rigorous as the paper it's based on. Specifically, it doesn't contain a precise statement of what Paxos accomplis...
Fluid Mechanics for Fun and Profit
Просмотров 1,1 тыс.7 лет назад
There are lots of fun applications of Fluid Mechanics! Extra-credit video for Eric Matthys' Fluid Mechanics class. Nov 17, 2004 Department of Mechanical Engineering University of California, Santa Barbara More videos, projects, code, presentations, contact info: justinppearson.com (c) 2004 Justin P. Pearson. All rights reserved.
Introduction to the Kalman Filter (Dynamical Neuroscience Symposium)
Просмотров 1,3 тыс.7 лет назад
This talk introduces the Kalman Filter an algorithm for efficiently estimating the state of a system from noisy measurements. Justin Pearson Jan 30, 2017 Introduction to the Kalman Filter Presented at the Dynamical Neuroscience Symposium University of California, Santa Barbara The Kalman Filter stirred up excitement in the aerospace industry in the 1960's as an intuitive, optimal, and easily-i...
How regenerative braking works
Просмотров 26 тыс.7 лет назад
Using a simple model for a DC motor, let's figure out how to use its momentum to recharge a battery. This was inspired by Phil Frost's post on the Electronics Stack Exchange: electronics.stackexchange.com/questions/56186/how-can-i-implement-regenerative-braking-of-a-dc-motor The code in this video can be downloaded here: github.com/justinpearson/How-regenerative-braking-works More videos, proje...
Pendulum + Beaglebone + Control Theory = Profit
Просмотров 11 тыс.7 лет назад
Pendulum Beaglebone Control Theory = Profit
Where's this signpost? A geodesy odyssey with Mathematica
Просмотров 2017 лет назад
Where's this signpost? A geodesy odyssey with Mathematica
IR camera, frying pans, and Mathematica
Просмотров 1,6 тыс.8 лет назад
IR camera, frying pans, and Mathematica
The UCSB Certificate in College and University Teaching
Просмотров 2888 лет назад
The UCSB Certificate in College and University Teaching
Example lecture: Java Programming: Exceptions
Просмотров 2238 лет назад
Example lecture: Java Programming: Exceptions
Mathematica: Sunset behind the UCSB Oil Rig
Просмотров 1568 лет назад
Mathematica: Sunset behind the UCSB Oil Rig
Thanks for publishing this video! Really insightful
Hi. I can't help but notice how smooth your voice is... Forgive me if that complement is inappropriate. But Also this helped me to go through my project as well ( even though I was tasked to do it in JAVA ) Thank you, sensei!
Amazing video for this topic! I came across this over 2 years ago when working on my Capstone project trying to implement regenerative braking for an E-Bike system. Unfortunately, the regenerative braking part of it didn't work out, but it stayed in my head as I went on to get my Master's in ECE where I wrote my thesis on the mathematical modeling of DC motor regenerative braking systems, which I just defended a week ago! I just wanted you to know that your video was a massive inspiration when I started learning about this. Thank you.
Love this, thank you!
Thank you so much for this. My probability class seems to think the homework is the best place to even introduce some of this notation. Feels like by the time I figure out what they're asking I could've already solved it. This is super helpful
Wonderful explanation and THANK YOU!
Amazing
Hello how would something like this work in python
This video is so great! I find the statistical notation provided in my classes to often be confusing, and often it is just because it is shorthand for something else that is more meaningful. I would love if you know of any more resources explaining various shorthand in probability and statistics. Also, I find the explanation you provide of random variables as functions to be very useful. One question I had was in regards to random variables in linear regression. The data points (X1, Y1), (X2, Y2), ... (XN, YN) are all composed of random variables. It seems confusing to think of these as functions. Instead, I have thought of random variables as "the output of a random process". I would love any clarification regarding this topic!
This video is very insightful. Understanding that math notation is really just shorthand for sentences in plain english can likely help alot of people afraid of math.
Which language is prefer for this project
Fresh air
the sounds of the marker makes this unwatchable.
I just got done taking a controls class. I loved it. However I really wish we would have had a lab. That would have been cool
Interesting demo lol
2:30 - Word "Algorithm" comes from his name :)
I tunes
Thanks for sharing this invaluable knowledge. Your explanation is very helpful.
Sweet.
Man this is great WE want more videos
wow, big thank you for this video finaly i found a good explanation & answers to all my questions... & i already consumed a lot of "content" on this
"The Silver bridge" point Pleasant W.VA
I'm doing a refresh of probability and stats, and the notations were killing me. Thank you for making this!
Gee Thanks. My head just exploded this morning..
What I'd like to know is if this regeneration system would charge the battery to full if it were to be continuously rotated by force? That is the only thing that is really important; because if it could; then it could charge the car while you are sleeping; if it were to be connected to a wind turbine that would be what could even be part of the vehicle. Because that is really what everyone want; free energy to run their car; that is everyone that isn't making money with selling energy.
Best explanation of Paxos I've seen. Thank you.
Nice explanation. Thanks.
The presentation was easy to follow and interesting. Well done brother 👍🏾
You summed up my Problem in the first instance!
Great presentation. I really liked his energy.
which algorithm did u use to find out the shortest path?
Overused 30 fps.
I've been mulling this over still more... It seems like there could be a scenario of "diminishing returns" in allowing such a close approach to steady state before switching over to battery charging; is there not likely a discoverable optimum rate at which to admit the charging pulses, such as may arise in triggering them before steady-state is so close at hand? It seems like you could make the pulses considerably more frequent before drastically shrinking them, while retaining the termination criteria to avoid reversal... This would admittedly reduce the braking torque but in a panic stop we're skipping regeneration anyway by holding the short, so these could be parts of a continuum... If we are doing this with FETs on an H-bridge with available PWM/CWG capabilities (and why would we not?), then it seems like just as we can modulate the driving voltage to provide reduced forward motion, surely we could modulate the braking in much the same way. Obviously we would need to provide a third state where after (1) the dead short provides braking and inductive current windup and (2) the north-going switch provides momentary current diversion to the battery, a state (3) would close both switches and provide coasting for the remainder of the anti-duty cycle. A bonus might be in that getting to this point seems to imply the potential for forward regeneration in case we happen to seek less than full operational speed, where a similar recharge pulse could be issued to release the inductive kickback at PWM frequency just before a coasting phase. If we're using FETs in a bridge, their anti-diodes are going to do this anyway when we turn the driving ones off after each pulse, so why not turn on the appropriate ones! Having laid out this rationale, it should be trivial to modify the drive orientation of the switches and just go on PWM-ing at the same frequency to accomplish all this, using current sensing to provide a trigger to the coasting state, with the state definitions being dynamic since there will be up to 4 different PWM paradigms including reverse (all of this can be accomplished with a PIC, including a hardware trigger to the coasting state at null current): For forward acceleration: (1) == upper-left and lower-right ON (drive). (2) == lower-left and upper-right-ON (regen). (3) == all switches OFF (coast) For forward deceleration: (1) == lower-left and lower-right ON (brake). (2) == upper-left and lower-right ON (regen). (3) == all switches OFF (coast). I have also been pondering how to surmount the current measurement problem for an empirical terminus of charging pulses, hopefully without sensing resistor. Perhaps a Hall effect sensor... Now I'll have to go and learn about that!
Great video! For a long time I was close to getting this, modeling motors as inductors and so forth but I couldn't account for inertia and a driven shaft. I never quite clicked into seeing that "back emf" is proportional to speed alone, and that it ultimately reverses the current in generator mode (I had been puzzled as to which top element should conduct for regen in an H-bridge setup). You video has fixed this for me; I am grateful! I assume that what you have analyzed here presents the means for maximum braking while capturing any available charge for regeneration. If I wanted to modulate braking torque, my presumption is some dead time should be injected after the upward pulse (upon cessation of charg flux) and before restoring the braking short via the lower switch, is this correct? Just trying to figure out where the fact that an open-circuited motor will coast fits into this whole paradigm so I can create any degree of braking up to the available maximum. I am trying to design a method to do this empirically. The target is a kids' ride-on car toy with a 6V battery and motor and the goal is to improve it just for giggles and to help get my daughter excited about this stuff. (Original setup is a pair of relays to operate the motor bidirectionally and if nothing else I want a soft-start/stop and speed modulation). I am using an H-bridge {(2x)SiRA99DP/(2x)SiRA90DP} and have ADC channels available. Can I simply and practically use a series sensing resistor and 2 kelvin-connected inputs to track the current, switching the bridge devices for regen on this basis alone? I can get 2 readings in about 50 microseconds so if those pulses are similarly short maybe I need more brute force logic like a comparator-based OR function for the high-side driver so I don't miss the cue and resume acceleration by disconnecting the battery too late? All things considered, I'm assuming that by transitioning my daughter's toy car to regenerative braking I can probably increase the longevity of a battery charge for her by at least 0.05%. I'm only integrating by eye, but it seems like very little of the invested energy can be recovered.
Can we please have a new way of writing and learning probability? Cause I'm fairly confident that no one is learning this stuff from first principles or axioms, and instead is just practicing until their black box in their head gets an intuition. We need to grow up, I think. Thanks for your video. I thought I was going crazy.
This is gold.
Your ahead of your time. Now recognized 9 years later as an artist. Thanks for the Kalman insights, its gold.
This is exactly what I was looking for. Thanks a lot for your effort!
Thank you so much for this! Stats becomes quite simple, the only problem is understanding notation - this helps a lot.
In my recommended anyone else 👍🏻
I just watched this to get some ideas about how to shake up online teaching! Do some flamboyantly bad teaching moves in a costume - that will wake the kids up!
Awesome presentation ✨
The probability is not as hard as notation. Would it be possible to include advance notations also ?
I guess there is a typo in the last equation at 17:49 (the expression for epsilon_i). The coefficient of K_i must be negative.
Yeah. You can see the error appear on the previous line, where the x - (...) from the line before that is expanded wrong. Note that the same typo appears in the formula for KF's measurement update in the slide at 31:00, but the code at 32:47 has it right with (I - K*C).
2222@@nylonsparrow
gold
wonderful explanation. keep up the good work
so litle energy you can generate back to the batrai?
Sweet!
where can I find the journals and scripts you use sir?
This is an underrated field.