HNHacker News
TopNewBestAskShowJobs

TFortunato

778 karma · joined January 23, 2011

I build robots.

Current: Staff Robotics Engineer at Chewy

Prior: Atlas @ Boston Dynamic; Robotics and Embedded Systems Engineering @ Amazon Robotics, Rethink Robotics, iRobot; Automation Engineering @ Ginkgo Bioworks.

Early Career: Electrical Engineer in Semiconductor Industry @ Applied Materials, Analog Devices, Allegro Microsystems

submissionscomments
TFortunato··on ODrive Project – Using inexpensive brushless motors in high performance robotics
That's a bit of an over simplification, but for a hobbyist application, sure.

You are absolutely right that the encoder is a BIG part of your system design, but the motor quality will make a difference. If you are doing something like a robotic arm, or slow speed motor control, you don't just care about encoder ticks, you also start caring about things like torque ripple in the motor if you want decent motion quality, regardless of how good your encoder is!

TFortunato··on ODrive Project – Using inexpensive brushless motors in high performance robotics
"Losing a step" makes me think you are thinking about stepper motors, which can be a problem, but for brushless motors it's a bit more complicated than that, since we don't drive them in the same stepwise fashion.
TFortunato··on ODrive Project – Using inexpensive brushless motors in high performance robotics
I'm on a cell phone, so I'll be brief, but can add more detail later. I've been doing motor drivers and embedded work (software and electronics) for robotics for a while, and did work on motor driver ICs before that, so this is kind of my wheelhouse! Briefly: I can tell you that the electronics and software are only a few parts of the overall equation here.

Yes, this does look like (from the demo video results), a decent driver design, however quality of the motor DOES matter, and will eventually become a limiting factor if you are lookig for high precision or high power, as you start to run into things like torque digging, motor heating, etc. There are also little things in the driver design that can bite you if you aren't careful, but I'd need to look at the schematics to comment.

I'm not trying to crap on the project or designer. There was clearly some thought and care put into this! I just don't want people to expect magic, or to think that the only difference between an industrial servo and a hobbyist motor is the electronics. That said, you can get VERY far with good electronics and control software!

A hobbyist project will probably do just fine with something like this (or a motor driver reference design from the chip manufacturers), and by the time they run into limits of those designs and hobbyist motors, they'll probably have a pretty good idea of where they need to go as a next step!

I'll take a closer look at this design later, and update if anyone is interested.

TFortunato··on Banks should let COBOL die
Depends on the government agency I'd say. For example: unemployement or social security checks going out, VA Hospital systems, etc.
TFortunato··on The U.S. wind industry now employs more than 100K people
No offense, but besides the advanced materials and mechanical design required to keep these things from tearing themselves apart, there is a lot more in the control and optimization software for these that goes into adjusting for changing weather conditions and power demand. I think your suffering from the classic engineer's disease of thinking something is a whole lot more simple than it is when looking at it from the outside.

Example article of some work going on in this space: https://www.google.com/amp/s/phys.org/news/2014-10-optimizin...

TFortunato··on React VR
On this topic... this is something I've wondered a bit about before (logical representations of 3d scenes -- interested as someone who comes from the robotics world, and curious how the people who did video games, VR, graphics tools, etc. thought about this stuff).

Obviously not the same use case, but I've tried looking before for good information on how scene graphs are structured and used, and haven't found much out there that was useful for a beginner who was interested in learning the details. If you have any resources or stuff you would be willing to share, or a well designed library that does it right (any language), I'd love to know about it!

Thanks!

TFortunato··on Apple has biomedical engineers developing sensors to monitor blood sugar
OK, so are you suggesting that a vegan diet or otherwise avoiding animal proteins will somehow prevent and/or cure juvenile diabetes then...

I give up, and suggest you learn to Google something more than just source to try to confirm (poorly!) your own biases about what diabetes is or isn't.

TFortunato··on Apple has biomedical engineers developing sensors to monitor blood sugar
Great, which still has nothing do with type 1 diabetes being caused by diet (or affluence). You can certainly manage weight gain or side effects in different ways, but the point still stands that Type 1 diabete is an auto-immune disease where the immune system attacks insulin producing cells in the pancreas. It is a very different disease than type 2 diabetes, where the pancreas still works, but the person has insulin resistance, and is the one people think of when they talk about diet and obesity leading to diabetes.

No amount of diet change is going to reverse the damage to the pancreas in Type 1 - it is not a "disease of affluence"

TFortunato··on Apple has biomedical engineers developing sensors to monitor blood sugar
I have to hand it to you for not even trying hard enough to read the parent's point about type 2 vs. type 1, and then linking to an article that mentions type 2 in the very first sentance.
TFortunato··on Tesla Passes Ford by Market Value
Maybe that's part of it. I'm sure another part of it is that Ford is already big enough, that there isn't a lot of growth potential there.

That said, it's not like they are exactly just making small refinements to gasoline engines, while SV companies are the only real innovators, chasing electric and self-serving vehicles. We just tend not to talk about it as much on HN (compared to Tesla, for example):

https://www.google.com/amp/s/electrek.co/2017/01/03/ford-new...

TFortunato··on Going Under the Knife, with Eyes and Ears Wide Open
Sorry you had to go through that (the waking up unable to breath).

FWIW, my son has had to go under general anesthesia almost a dozen times now, and the protocols that the anesthesiologists use at our hospital use now require a respiratory therapist to be in the room when anesthesia is removed, and it is done in stages, so that it is made absolutely sure he is breathing on his own before the breathing tube is removed and he is fully awakened. I'm sure part of it was driven by the fact that what you went through is both awful, and should have been prevented.

I'm saying only to let you know that I really empathize with the anxiety it can cause, and that if you ever are in need of surgical care again, perhaps you should discuss it with your surgeon or care team before-hand..hopefully they have protocols in place where you won't have to go through that again.

On another note...holy crap on the endoscopy! I can't imagine how much local anesthesia that must have taken, or how odd that must have felt!

TFortunato··on Going Under the Knife, with Eyes and Ears Wide Open
I too have had mine removed under only local anesthesia, and fairly recently at that (in my early 30's). It's interesting to see some people say it was a horror show while others said "It wasn't so bad", since I've experienced both!

The reason why I finally had my wisdom teeth removed, was that I had one of them in my lower jaw, that was growing in almost sideways into the molar in front of it, causing a large abcess in the tooth. My dentist was surprised it had never caused me pain before this (I had spent a few years bouncing around, and hadn't gone to the dentist for years, so the fact it got so bad was likely entirely preventable...)

Anyways, because of the way it had grown pushing into the other tooth, the roots of the wisdom tooth were very curved (not quite fishhooks, but close!), so the procedure took a couple hours, while under only local anesthesia, where I could hear and fell him drilling, and chiseling, and digging in my jaw to remove the tooth in pieces. At one point in the middle of the procedure, he needed more images to see what was left, and left large metal pins in the holes from my roots as reference points of how far he had gone into my jaw so far. He gave me the warning "You really don't want to bite down while these x-rays are done", hah! Needless to say, when all was said and done, this was a pretty painful recovery, and still had bits of broken tooth root and bone pushing out of my gums for a few weeks after.

On the other hand, when I removed the wisdom tooth from the opposite side of my jaw, to keep things even, it had grown in straight, and was a very simple procedure where the tooth came out in one piece, and I had very little pain in recovering.

Everything had healed up beautifully now, and the dentist was able to save the abcessed molar, but man, I don't want to go through that again ever, and I can total understand why some people elect general anesthesia for such a procedure, since it really is a crap-shoot on how "simple" of a procedure it is. General may have its risks, but I'm sure people being scared to ever see their dentist again comes with risks as well!

TFortunato··on It's Better To Animate a Film Alone, Even If It Takes Years
To play devil's advocate, I think what was suggested is that for the artists who do this kind of work, the eschewing of using a computer IS what makes this, if not superior, at least their preferred method of working. For at least some artists doing this work, the whole point is that everything you see is drawn by hand because they want to do that work, not for lack of a better tweening algorithm, machine learned or otherwise.
TFortunato··on Python to Circuits
On this note, as an embedded systems guy, I was happy to see STMicro thinking about tools like this, with the STCubeMX software they've been developing: http://www.st.com/en/development-tools/stm32cubemx.html

One of the pains about larger embedded designs, is that while newer processors/microcontrollers have so many different peripherals and ways to use them, there are just as many constraints about which functions can go to which pins, how each peripheral is clocked, etc..

Their tool is similar to what you mentioned, in that you tell it what peripherals you want to use, and how, if you have any pins you need to lock down, etc. It will then solve these constraints and give you a pinout and clock setup that will work, along with boilerplate code to get you started.

I would love to see an open-source effort in this space...basically a parametric framework / boilerplate generation for embedded systems designs (hardware and software).

TFortunato··on Mo.js: motion graphics toolbelt for the web
Out of curiosity, are you running an Nvidia card? (This may apply to radeons as well, but I don't have any experience)... For the NVidia drivers at least, there is an option to select which graphics card to use on a per-application basis. Normally the drivers default to auto-selecting what it thinks is best, but you can override that in cases like this.
TFortunato··on Neuroscientists find long-sought bat sat-nav neurons
This was a feature of the GameShark for Nintendo 64 as well for finding new codes beyond what they published.

For example, if you had 100 rounds of ammo left, you would press a button on the cartridge, and have it search the ram for the value 100. Drop back into the game, fire off a shot, and hit the button again to search for which of those values changed to 99. Repeat until you found target area in memory and make a code for it.

The device also had a parallel port interface to talk to it via a pc, but I remember less about using that when I was a kid. Pretty basic technique, but as a kid interested in technology, it's stuck with me to this day!

TFortunato··on Georgia Tech Offers Online Master of Science in Analytics Degree for Under $10K
Question for you if you don't mind...

I've been out of undergrad and working in industry for close to a decade as a Software Engineer and/or Embedded Systems Engineer. I feel I'm doing pretty well in my career, but I've been looking at the online masters in CS as a way of showing that I'm dedicated to continuing learning, and to maybe open up some new doors for myself down the road.

From your perspective as a hiring manager, would this be worth the time and effort? It's not like I wouldn't be interested and learning new stuff anyways, but if I'm going to go invest the time and money to do the degree vs. learning on my own, it would be nice to know if it was worth anything.

(No sweat if you don't feel like answering, or want to take this private. Thanks!)

TFortunato··on Negative Resistance with a Single Atom
From my background in EE, I can tell you a bit about the why at least. While I have some background in device physics, it's as an EE not as a physicist, so I probably won't be as helpful on the how.

- For analog type devices, negative resistance gives you a way to easily make oscillators, and negative feedback loops. Practical applications could be things like smaller or lower power wireless chips, or more accurate filters / oscillators in those chips. Useful for everything from wireless chipsets to things like google's project soli.

- For digital type devices, the big draw is that you can make a memory cell (aka flip-flop) from one active component instead of 2. This could potentially halve the size of some of the digital logic in your microprocessor (at least the registers and all the temporary storage used in your processing pipelines!)

- Talking more pie-in-the-sky, some models of neurons include negative resistance areas...analog neural nets on a chip using these will probably support a grad student or two as a research project!

Start reading here: https://en.m.wikipedia.org/wiki/Negative_resistance

TFortunato··on Ask HN: Why are flash memory prices going down so much faster than RAM?
I mean, 5 different companies have pleaded guilty to price-fixing in the DRAM market, so I don't think what is suggested here is too controversial.

https://en.wikipedia.org/wiki/DRAM_price_fixing

TFortunato··on Predictive Learning [pdf]
Re Westworld: The theory of consciousness explored in the show is explored in more detail in Jaynes' The Origin of Consciousness in the Breakdown of the Bicameral Mind, as alluded to both in the show and in the title of the final episode. I've just picked it up, and it's a pretty interesting read so far. I've also noticed that a lot of little details from the book used in the show, such as referring to memories as "reveries" at points, and talking of minds as "hosts" of consciousness. I may need to re-watch the show after finishing the book!

I found a pdf version of the book here if you are interested. http://selfdefinition.org/psychology/Julian-Jaynes-Origin-of...

TFortunato··on Reinforcement Learning with Unsupervised Auxiliary Tasks
Exactly what I was thinking, and put much more succinctly!
TFortunato··on Reinforcement Learning with Unsupervised Auxiliary Tasks
Quick reply to my own comment. But another reason it seems this is helpful is that in many reinforcement learning tasks, like games, rewards are few and far between, so these goals also give the agent oncentive to keep trying new things / learning about the environment / task in the presence of sparse rewards?
TFortunato··on Reinforcement Learning with Unsupervised Auxiliary Tasks
Interesting. I'm not a deep learning guy, but from what I can gather, the new auxiliary tasks are to be rewarded for "pixel changes" and "network features".

I haven't nearly finished reading the paper, but is it safe to say this is similar (at a very high level), to a type of "novelty search", where the agent is searching not only for a policy that is directly accomplishing the task at hand, but also for novel stimulus (in the case of pixel changes), and novel internal states (features, or maximally activated hidden nodes in the language of the paper), and that the benefit of this would be to more easily find relevant features that could be useful in the "big picture" task, and maybe not get as stuck in a non-optimal policy?

(I may be understanding this completely wrong...just an embedded guy looking to get more into this world, haha)

TFortunato··on Show HN: Interactive deep convnet visualization for Keras and Tensorflow
Do you have a link to the paper you are referring to?
TFortunato··on Ask HN: Ever been hired through a “who's hiring” post?
Hey I was going to post about iRobot, and you beat me to the punch again!

On this - I was one of those hired by iRobot from one of these threads, and this was the guy who I first spoke to about the job!

TFortunato··on Twitch Plays Baxter
This is awesome. I was about to ask at the office who set this up, but I see you bought it second hand.

Still plan on sending to everyone at the Rethink office to check out!

TFortunato··on Cartographer – A real-time simultaneous localization and mapping library
Full disclosure: former iRobot engineer speaking -

Coincidentally, for the latest Roomba (the 980), we did have a camera on the robot which did visual SLAM for precisely that purpose...the idea being it would map your house while it went, and if it needed to stop for a recharge, it can go back to its dock for a recharge, and then resume cleaning right where it left off.

https://www.technologyreview.com/s/541326/the-roomba-now-see...

We also worked on a remote presence (teleconferencing) robot, which used SLAM to map the building it was in, so that a user could just click on a location on a map or schedule the robot to be in a conference room for a certain time, and the robot could autonomously drive there without someone remotely controlling it. It's a very cool technology that has a lot of really useful, non-scary, applications!

TFortunato··on A SpaceX Falcon 9 rocket has exploded at Cape Canaveral
Thanks for the info! My experience in the past was mainly with large LN2 dewars, with liquid helium refrigerators added to re-condense the N2 gas back to a liquid, limiting the pressure (and limiting how often you needed and LN2 refill!). We had blow-off valves as a fail safe if pressures got too high if the cooling head ever failed. I'm sure working with cryogenic fuels is a whole different world.

As I commented, it's not that there aren't cryogenic temperature pressure relief valves, just that they may not be what you want!

TFortunato··on A SpaceX Falcon 9 rocket has exploded at Cape Canaveral
I can't think of the manufacturer of the top of my head, but there are definitely off-the-shelf options. They are used on the LN2 tanks and dewars you see at a lot of electronics (and I'm sure other) manufacturers, and I'm pretty sure are standard kit for any cryogenic vessel. I think the bigger issue, as one commentor alluded to, is that you might not want to vent whatever is in the rocket straight to atmosphere.
TFortunato··on Teaching AI to race using NNs and genetic algorithms
From what I can see, the course you mentioned looks very cool, however it seems a lot of the slides and linked papers are password protected. Do you have copies mirrored and/or know if the professor would be OK with these being shared?

(Don't worry if not, As much as I'd love to browse some of whats on there, I don't want to make extra work for you or get anyone in trouble!)

← PreviousPage 8 of 9Next →