Beautiful aesthetics aside, functionally it is one of the most impressive and polished PC based robots I’ve ever seen as well; custom software that handles object recognition, mapping and navigation, as well as voice commands, just to name a few. If all of that wasn’t impressive enough, creator Dave Shinsel has the schematics, documentation and even source code available on his site! Definitely check out Dave’s entire site, he has an impressive body of work, and his Youtube Page has some very cool videos demonstrating Loki in action.

XMOS event driven processors allow you to execute code in parallel which could open up a number of possibilities to evolve your robots “mind” to the next level!

I have some development kits to give away potentially, but before I give away all the details, I just wanted to see if there was any interest.

People have made some cool stuff with our tech before (See videos).

Some details of the processor itself which will be on the dev kit (to be released soon):

- Single core device (Although we do have quad core versions – ask me)
- 400 MIPS per core.
- 8 Threads per core.
- 64Kb RAM
- 8KBytes OTP memory for applications, boot code or security keys, with security mode
- 64 user I/O pins
- Support for high performance DSP (32 x 32 → 64bit MAC) and cryptographic functions
- Time aware ports provide up to 10ns timing resolution
- Designs implemented using a software-based design flow (can program in C or XC – very similar to C but with support for extras such as parallel execution)
- Scalable – can connect many kits together for crazy amounts of processing power using Xlinks.

Many roboticists are well aware of Roborealm and all of its awesomeness, so why am I preaching to the choir here? Because they recently published a tutorial on using vision for obstacle avoidance, and one look at it had my jaw dropping. Some very useful stuff here, as well as downloadable example files and code showing how to use it in a real world application. Anyone interested in machine vision and PC based robotics should definitely check this out! While Roborealm is available for free the cost associated with its development is by no means cheap, so if you enjoy the product please consider donating! Information on donating to Roborealm can be found on their site.

Here is my quick review here on the forums of this camera:

Image quality is mediocre, but good enough to view things at a max of 8-10 feet. Frame rate is 30+ and response time is in the 50ms range running the camera at 320×240.

Main board is 10cm x 5.5cm. Camera board is 5.5cm x 3cm and attaches to the main board at a right angle. It appears to be a 1mm spaced header pin port, so it could easily be extended for custom mounting solutions. Entire camera unit stripped down and including antenna weighs 91 grams. The antenna is a u.FL connection, so this could be replaced with a custom solution if needed.

Overall a solid little camera for the price. If you’re on a budget, this is about as good as it gets. Keep in mind for ‘competition’ performance, you’ll need to run it at 320×240, but that can be stretched. Image quality is enough to pilot and aim I would say, but I wouldn’t expect to read text with it. Definitely usable for Mech Warfare within 10-15 feet though. Oh and the IP web interface uses Java and/or Active X, so you could pull the control from it and add it into a custom form easily.

Here are some pictures! (click to enlarge)

Since we don’t care about the bulky exterior casing, here’s what it looks like stripped down:

And this is with the camera board removed:

I’m sure this can’t last forever, so make sure to check it out while it’s active!

Four human-controlled rows, four robotic ones. At the right, you can see the custom PIC-based servo controller board.  Let’s take a closer look at the drive mechanisms:

Recognize that actuator?  That’s a Robotis AX-12 Dynamixel.  It’s the cornerstone of the wildly popular Bioloid system, and it’s one of our favorite robot servos on the market.  These actuators can be set to servo mode or continuous rotation mode.  In CR mode, you can get accurate positional feedback.  What you end up with is functionally the same as a highly accurate, powerful, serially controlled stepping motor; at a fraction of the cost.

Here’s the whole rig.  As you can see you have the table, a camera watching the table, a computer processing the video feed and sending commands to the actuator controller board, and actuators pulling and twisting the handles.

The team wrote their tracking software, which tracks the ball and maneuvers the players accordingly, in Java.  Matlab was considered, but the plan was scrapped because Matlab is too resource-hungry.

The custom actuator control board is pretty cool.  Two different types of outputs for the two styles of servos.  The kicking motion is handled by standard PWM hobby servos, while the lateral motion is handled by AX-12 Dynamixels.

Project home page
Found via Hack a Day

All images belong to the design team.  They were clipped from the project’s final report document or captured from the video.