Enigma neural network

Decoding Enigma Using A Neural Network

[Sam Greydanus] created a neural network that can encode and decode messages just as Enigma did. For those who don’t know, the Enigma machine was most famously used by the Germans during World War II to encrypt and decrypt messages. Give the neural network some encrypted text, called the ciphertext, along with the three-letter key that was used to encrypt the text, and the network predicts what the original text, or plaintext, was with around 96-97% accuracy.

The type of neural network he used was a Long Short Term Memory (LSTM ) network, a type of Recurrent Neural Network (RNN) that we talked about in our article covering many of the different types of neural networks developed over the years. RNNs are Turing-complete, meaning they can approximate any function. [Sam] noticed the irony in this, namely that Alan Turing both came up with the concept of Turing-completeness as well as played a big part in breaking the Enigma used in World War II.

How did [Sam] do it?

Continue reading “Decoding Enigma Using A Neural Network”

Rushing The Design And Construction Of LED Centerpieces

‘Dragon Flame’ RGB LED table centerpieces, by [Alex Lao]
Sometimes the most important thing is getting something done.

[Alex Lao] was recently in such a situation. His sister was getting married and he designed, built, and delivered twenty RGB LED table centerpieces in a rush. There were no prototypes made, and when the parts arrived all twenty were built all at once over a single weekend. These table centerpieces are illuminated by RGB LEDs and battery-powered, but have an option to be powered by a wall adapter.

[Alex] helpfully shared some tips on reducing the production risks and helping ensure results in such a limited time frame. His advice boils down to this: reduce the unknowns. For Alex this meant re-using code and components from a previous project — even if they were not optimal — so that known-good schematic and footprint libraries could be used for the design.

From one perspective, the PIC32 microcontroller inside each lamp is overkill for an LED centerpiece. From another perspective, it was in fact the perfect part to use because it was the fastest way for [Alex] to get the devices working with no surprises.

For an added perspective on needing to get production right the first time on a much larger scale, be sure to check out getting an installation made up of 25,000 PCBs right the first time.

We Should Stop Here, It’s Bat Country!

[Roland Meertens] has a bat detector, or rather, he has a device that can record ultrasound – the type of sound that bats use to echolocate. What he wants is a bat detector. When he discovered bats living behind his house, he set to work creating a program that would use his recorder to detect when bats were around.

[Roland]’s workflow consists of breaking up a recording from his backyard into one second clips, loading them in to a Python program and running some machine learning code to determine whether the clip is a recording of a bat or not and using this to determine the number of bats flying around. He uses several Python libraries to do this including Tensorflow and LibROSA.

The Python code breaks each one second clip into twenty-two parts. For each part, he determines the max, min, mean, standard deviation, and max-min of the sample – if multiple parts of the signal have certain features (such as a high standard deviation), then the software has detected a bat call. Armed with this, [Roland] turned his head to the machine learning so that he could offload the work of detecting the bats. Again, he turned to Python and the Keras library.

With a 95% success rate, [Roland] now has a bat detector! One that works pretty well, too. For more on detecting bats and machine learning, check out the bat detector in this list of ultrasonic projects and check out this IDE for working with Tensorflow and machine learning.

You Know You Can Do That With A 555

Hardly a week goes by that we don’t post a project where at least one commenter will lament that the hacker could have just used a 555. [Peter Monta] clearly gets that point of view. For a 555 design contest, he created both digital logic gates and an op amp, all using 555 chips. We can’t quite imagine the post apocalyptic world where the only surviving electronic components are 555 chips, but if that day were to come, [Peter] is your guy.

Using the internal structure of the 555, [Peter] formed a basic logic gate, an inverter, latches, and more. He also composed things like counters and seven-segment decoders. He had a very simple 4-bit CPU design in Verilog that he was going to attempt until he realized it would map into almost 400 chips (half of that if you’d use a dual 555, but still). If you built this successfully, we would probably post it, by the way.  You can see a video of the digital logic counter, below.

Continue reading “You Know You Can Do That With A 555”

Handheld Gimbal With Off-The-Shelf Parts

For anything involving video capture while moving, most videographers, cinematographers, and camera operators turn to a gimbal. In theory it is a simple machine, needing only three sets of bearings to allow the camera to maintain a constant position despite a shifting, moving platform. In practice it’s much more complicated, and gimbals can easily run into the thousands of dollars. While it’s possible to build one to reduce the extravagant cost, few use 100% off-the-shelf parts like [Matt]’s handheld gimbal.

[Matt]’s build was far more involved than bolting some brackets and bearings together, though. Most gimbals for filming are powered, so motors and electronics are required. Not only that, but the entire rig needs to be as balanced as possible to reduce stress on those motors. [Matt] used fishing weights to get everything calibrated, as well as an interesting PID setup.

Be sure to check out the video below to see the gimbal in action. After a lot of trial-and-error, it’s hard to tell the difference between this and a consumer-grade gimbal, and all without the use of a CNC machine or a 3D printer. Of course, if you have access to those kinds of tools, there’s no limit to the types of gimbals you can build.

Continue reading “Handheld Gimbal With Off-The-Shelf Parts”

Retro DEF CON Badge Made From ’80s Parts

DEF CON 25’s theme was retro-tech, and [xres0nance] wasn’t kidding around in the retro badge he built for the convention. The badge was mostly built out of actual parts from the ’80s and ’90s, including the perfboard from Radio Shack—even the wire and solder. Of the whole project just the resistors and 555 were modern parts, and that’s only because [xres0nance] ran out of time.

[xres0nance]  delayed working on the badge until his flight, throwing the parts in a box, and staggering to the airport in the midst of a “three-alarm hangover”. He designed the badge on the plane, downloading datasheets over in-flight WiFi and sketching out circuits in his notebook.

The display is from an old cell phone, and it uses a matrix of diodes to spell out DEFCON without the help of a microcontroller. Each letter is powered by a transistor, with specific pins blocked out to selectively power the segments. He used a shift register timed by a 555 to trigger each letter in turn, with the display scrolling the resulting message.

We publish a lot of posts about con badges. See our DEF CON 2015 badge summary for a bunch of badges that we encountered at in Vegas.

ZBeam

Hackaday Prize Best Product Finalist: Shape Shifting Structures For Space

While [Elon Musk] and [Jeff Bezos] are working on getting us to Mars and the Moon, [Ronald Jaramillo] is working on building structures once we get there. To that end, he’s been developing the ZBeam, two rolls of links that zip together like a zipper to form a rigid beam.

ZBeam making, regolith munching machine
ZBeam making, regolith munching machine

Initially stored in a compact cube targeted to eventually fit in a CubeSat’s dimension’s, 100 mm x 100 mm x 100 mm, the beam emerges from within the cube and will be able to connect with other cubes to form rigid structures. His hope is that they can one day be made automatically from lunar or Martian regolith (loose surface dirt) munching machines. His current one has 160 mm sides and uses a servo hacked to turn continuously.

In his hackaday.io project logs he shows the trial and error he’s gone through to get to his current stage: experimenting with the links to form a more rigid beam, fine tuning the unreeling of the rolls of links to prevent jamming, adding a safety-ratchet-gear to the gearing to overcome speed issues, and more. He currently 3D prints as many connected sets of links as he can on his Prusa i3, and then manually connects sets together to make a longer chain, but he has his eye on the Printrbot Printrbelt for printing arbitrarily long chains in one piece.

You can see one pretty impressive iteration of the ZBeam in action in the video below and more is on his project page. In fact, the judges for the 2017 Hackaday Prize liked [Ronald]’s projects so much that they designated it as a Best Product finalist.

Continue reading “Hackaday Prize Best Product Finalist: Shape Shifting Structures For Space”