Easy Camera Tracking With A Quadrocopter

January 4, 2012 by Brian Benchoff 11 Comments

[DJ Sures] has been pulling all-nighters lately to get his AR Drone Parrot build off the ground. Now that it’s up and flying around, he managed to get it to follow objects around the room using on board cameras.

For the build, [DJ Sures] used the AR Drone ‘flying video game’ quadrocopter. This toy has two on board cameras that can viewed over wifi. All that’s needed is some interesting software to make things fun. The camera tracking of EZ-Builder software was brought into the mix so the AR Drone can be controlled via object or speech recognition, wiimotes, tablets, or terminals.

[DJ Sures] has come up with some ~~slightly terrifying~~ awesome builds like a Bluetooth Teddy Ruxpin, realistic Wall-E, and an awesome Omnibot 2000 refurb. This is his first flying hack, and the first to fully exploit the camera tracking of the EZ-Builder software. Check out [Sures]’ copter following him around a room after the break.

Continue reading “Easy Camera Tracking With A Quadrocopter” →

Neural Networks Control A Toy Car

January 3, 2012 by Brian Benchoff 27 Comments

After taking the Stanford Machine Learning class offered over the Internet last year, [David Singleton] thought he could build something really cool. We have to admit that he nailed it with his neural network controlled car. There’s not much to the build; it’s just an Android phone, an Arduino and a toy car. The machine learning part of this build really makes it special.

A neural network takes a whole bunch of inputs and represents them as a node in a network. Each node in [Davids]’s input layer corresponds to a pixel retrieved from his phone’s camera. All the inputs of the input layer are connected to 64 nodes in the ‘hidden layer’. The nodes in the hidden layer are connected to the four output nodes, namely left, right, forward and reverse.

After training the network and weighting all the connections, [David] got a toy car to drive around a track. Weird, but it works. All the code is up on github, so feel free to take a look behind the inner machinations of a neural net. Of course, you could check out the video of [David]’s car in action after the break.

EDIT: We originally credited [icebrain] as the author. Our bad, and we hope [David] doesn’t hate us now.

Continue reading “Neural Networks Control A Toy Car” →

Easy Rotary Encoding For Your Projects

January 2, 2012 by Mike Szczys 10 Comments

Want to monitor how much a wheel has turned in your project? Then you need a rotary encoder! Here’s a way to add rotary encoding without changing the mounting method of your wheels (translated). [Jorge] added it as a way to improve the functionality of this line-following robot. It uses a paper encoder wheel which is monitored by an optical sensor.

The paper wheel consists of alternating white and black pie pieces. You can make this with a felt-tipped marker, or use a tool like the one we featured a couple of years ago to print out a disc rendered to your own specifications. This is glued to the inside of the wheel and monitored by a CNY70 reflective sensor (the same one used in that electric keyboard retrofit).

The homemade board which holds the sensor can be seen mounted on top of each wheel’s motor. It requires three wires, voltage, ground, and data. The data line is connected to the output of the phototransistor in the CNY70 package so it can be used with a microcontroller interrupt for easy integration with the firmware driving the robot.

[Jorge] goes into some detail about how the added data helps to improve the speed performance seen in the clip after the break.

Continue reading “Easy Rotary Encoding For Your Projects” →

Zombie [Jackson] Bot Dances To Thriller

December 31, 2011 by Mike Szczys 2 Comments

[Marc Cryan] built this little bugger which he calls Wendell the Robot. But what good is an animatronic piece like this unless you do something fun with it? That’s why you can catch the movements matching [Michael Jackson’s] choreography from the music video Thriller in the clip after the break.

This is a ground-up build for [Marc]. He started by designing templates for each of the wood parts using Inkscape. After printing them out, he glued each to a piece of 1/4″ plywood and cut along the lines using a band saw. We don’t have a lot of adhesive spray experience, but he mentions that the can should have directions for temporary adhesion so that the template can be removed after cutting.

During assembly he makes sure to add servo horns for easy connection when adding the motors. All together he’s using five; two for the wheels, two for the arms, and one for the neck. A protoboard shield makes it easy to connect them to the Arduino which is used as a controller.

Continue reading “Zombie [Jackson] Bot Dances To Thriller” →

A Little Tweeting “Ewok”

December 27, 2011 by Jeremy Cook 22 Comments

Tweet Receiving, that is. This Ewok model, named “Ewen the Cheerlight,” is able to rotate its head left and right as well as show expressions. The most interesting feature of this hack, however, is that the little Ewok actually wakes up each time one tweets a “colour” to @cheerlights and lets it’s owner [Joel] know what he thinks of the “color” you’ve chosen. [Joel] insists that he’s like this featured on [HAD], although it remains to be seen if it will soon be turned off after the tweets start rolling in…

As far as how the device works, the head is turned with a simple hobby servo motor, and the expressions are shown on a LED matrix. The model itself is built from a polystyrene ball and an old table lamp. The build really looks awesome.

All of this is controlled by an Easy USB Interface Board which is listed on [Pozible], the Australian version of [Kickstarter]. Be sure to check out the video of “Ewen” in action after the break! Continue reading “A Little Tweeting “Ewok”” →

A Robot… That Can See Through Walls!

December 26, 2011 by Jeremy Cook 14 Comments

Robots on four wheels are fun on their own merits, but one thing that most lack is the ability to see through walls. With it’s onboard radar system, this bot is equipped to see objects that a person couldn’t normally detect on the other side of the wall.

Although some of the more “nuts and bolts” details of this build are missing, the robot uses an Ultra-Wideband Radar system called the [D1] Radar System. This system can, according to their documentation, “Avoid false positives caused by vapor, dust, smoke, rain or other small particles.” Apparently this means drywall as well if programmed correctly.

In the video after the break, the robot’s sensor package is programmed to ignore anything within 1.5 meters. This allows the robot to mirror the movement of the apparent shelving unit on the other side. This sensor could certainly have some interesting robotics applications besides imitating a rolling shelf, so we’re excited to see what it will be used for!

Continue reading “A Robot… That Can See Through Walls!” →

Tribot Robotics Platform

December 19, 2011 by Brian Benchoff 5 Comments

For a number of children born of geek parents, the WowWee Tribot is sure to make an appearance underneath a Christmas tree this year. By New Year’s, though, this toy will surely make its way to the back of a closet to sit unused until spring cleaning. It’s a shame to let such an interesting robotics platform go to waste, so [haltux] sent in a nice guide to unlocking the motor controller of this talking robot.

The ‘legs’ of the WowWee Tribot have three omnidirectional wheels mounted 120 degrees apart. We’ve seen this drive system before, so getting a pre-built platform out of the toy box is pretty interesting.

[haltux] found three H-bridges inside the Tribot and connected the direction and enable pins for each motor directly to an Arduino. The build was a success, and the new robot platform scurried along the floor. There are also rotary encoders on the Tribot, but these run at 12 Volts. [haltux] said he’ll cover these in a future post, and we’re waiting to see it.