Huey
@ January 17, 2011
- Actuators / output devices: Hitech HS-81 servos and Tamiya twin motor gearset in low gear
- Control method: Xbee wireless controller, autonomous or remote over C# application
- CPU: Atmel ATmega328
- Operating system: Arduino C
- Power source: 6 AA batteries power everything
- Programming language: C and C#
- Sensors / input devices: Parallax Ping
- Target environment: home
-> Video now added <-
This robot, Huey, improves structurally on my previous design, providing a more finished look and a much more durable chassis, but the general approach and the technologies involved are the same. I ditched the Tamiya track and wheel set, but kept the Tamiya twin motor gearbox as it worked admirably (aside from the EMI issues caused by the low voltage motors) and that price is just so appealing.
I bought some miniature track links and pins from Pololu and used some right-angle aluminum for rigidity. The body was built using “project enclosures” from Radio Shack, and the neck pan and tilt are made from very slightly modified servo brackets from Lynxmotion.com. I figured out it’s not cheaper to build your own servo pan and tilt system if you value your own time!
Huey supports remote control and autonomous modes as before, although Huey adds support for triggering autonomous mode through a momentary button on his back. The remote wireless connection is provided by Xbee wireless antennas, which provide a COM connection between a PC and the Arduino unit inside the robot. I wrote a piece of software in C# that runs on my laptop which monitors key down and key up events on the keyboard. The “WSAD” buttons on the computer keyboard provide guidance similar to that found in a FPS game. Power to the two Tamiya motors is determined at the laptop using a C# slider control, with power levels being transmitted to the robot at each key press.
I improved the behavior of the autonomous mode considerably since the previous design, and I’m happy to report I’m still only using about half of the available Arduino memory for my sketch. Although obstacle detection is working much better using Ping alone in this model, there is still no cliff detection, and the Ping sensor is prone to the occasional false-positive reading when put in the EMI-noisy environment caused by those 3V Tamiya motors.
This model has 8 AA batteries on board as compared to the previous 6. This increased power is still within the range of the motor controller from LadyAda that powers this robot, but heat was building up on the L293D H-bridges, so I had to add some heat sinks because I was hitting the built in thermal shutdown during heavy use.
If you'd like to see more photos, please look here, and thanks for your interest!
@ Mon, 2011-01-24 04:42
looks very neat
have you design a logo for him?
@ Mon, 2011-01-24 16:49
Hi digi01, I love your new
Hi digi01, I love your new frisbee bot, your snake, and your other stuff too. Great craftsmanship!
I haven't thought about a logo, do you have any suggestions? I was thinking about calling this line of robots "DDDOT" as a joke, standing for "DDDOT Doesn't Drive On Tracks".
@ Tue, 2011-01-18 11:06
Nice looking bot, perhaps a
Nice looking bot, perhaps a little top heavy?
You said the button on the back was for selecting between autonomous and remote control. Have you considered using the Xbee's line passing feature? This would allow another button on the transmitting Xbee to do the same job. Then you can change modes remotely as well as locally.
@ Tue, 2011-01-18 22:41
Thanks OddBot, I love your
Thanks OddBot, I love your work BTW!! Much respect.
As the designer of the incredibly awesome Dagu Thumper, which I've researched a little and lusted after greatly, I can see how you might see my low priced piece of junk as top heavy. There is actually very little weight up top, with the batteries in the bottom it's most of the weight is down low, but with this bot, which is part of a series of bots, my goal is to explore emergent group behavior, if I can muster the time and the brains to do it. I don't mind if it can only handle fairly gentle inclines, as this round of work is more about behavior and robot to robot communications than anything particular about the chassis. I'm trying to keep the unit cost as low as possible without going so cheap that sensor readings are impacted, which was why my first bot failed.
BTW my third bot, Dewey, uses the good old RP5 chassis with the same RadioShack project enclosure screwed to the top. I'm starting to figure out I want to spend time on software and not chassis fabrication. If you think there is a better approach, please guide me!
Regarding the button, what you describe is almost exactly what I did, but mostly using software. I'm hoping to put up a video tonight that will demonstrate. There are two ways you can initiate or deactivate remote mode, which is just a true/false state. You can press the momentary button, which is attached to a hardware interrupt on the Arduino, one press initiates remote mode, another press deactivates. On the software side, a press of the button "Q" on the PC keyboard is passed via Xbee through the COM link and initiates/deactivates the same code as the interrupt. (Why the letter Q? It's close to "ADWS" on the keyboard.)
@ Wed, 2011-01-19 04:35
Junk Bots
I would never call a robot a piece of junk. Look at all my early robots on LMR. Breadboards, wires and hotglue with mecanno for structure.
I was just thinking your bot needs a bigger wheel base.
@ Wed, 2011-01-19 06:33
You're point is well
You're point is well taken. I would never call someone else's work junk either. It's all about the love of creation, right? And perhaps we are more hard on our own work than we are on the work of others.
You're right about the wheel base, it's pretty tippy on even a moderate incline. I guess you gotta pick your battles. By the way, I know you've done a ton of great stuff, but I really love your compound eye work, just fantastic. My third bot was somewhat inspired by that, I'm measuring distance via Sharp IR and Ping, then examining the results to try and detect false positives and lousy readings. That work is leading me to believe ping sensors are generally troublesome.
@ Wed, 2011-01-19 15:53
I want to make an ultrasonic
I want to make an ultrasonic version of the compound eye but the problem is that as a product it might be hard to keep the price low. The IR eye is cheap enough for anyone to afford.
Most people think the range is limited to about 20cm ( 8 inches ) but by changing just 2 resistors you can easily double the range. The downside is if you make a mistake in your code then you can fry the IR LEDs.
The range is set to "Safe Mode".
@ Mon, 2011-01-17 06:08
Nice- looks clean, hope to
Nice- looks clean, hope to see a video. BTW, I ditched my Tamiya FA-130 3V motors for Solarbotics RM3s running around 6 volts and have been very pleased with the results in my tamiya twin motor gearbox.