I looked at the project FIGARO. The project is a six-leg robot for cutting lawn, painting and even dancing and playing piano. As the video clip showing, the FIGARO got a high exposure to the publicity in Hungary. It is fun to look the FIGARO playing all the skills and it looks complicated. I noticed the FIGARO pauses in moving each step in the real time video (the video clip shows most time by playing back in higher speed rate). I would like to see why. So I checked the hardware and software documents posted on FIGARO page. To my surprise, the system, as well as the documents, is very simple. The software diagram seems a simple flowchart. I cannot even read the characters from the flowchart (it may be in Hungarian?). The software source code is very simple, just one .c file and a couple .h files. By putting all the pieces of the puzzle together, I finally got the picture. At each step, the marker sends a radio trigger to the FIGARO followed by the ultrasound pulses. When the FIGARO receives the radio signal coming from UART2, it stops all the activity and waits for the ultrasound signal. By measuring the time delay of the ultrasound arriving the FIGARO figures out the distance between the marker and itself. The six legs control is through the UART1 port. The PIC32 in this project is playing a simple role in repeated steps: it reads a preprogrammed coordinate parameter from a file in the SD card; corrects the coordinate by getting time delay from the markers and sends the movement command to UART1 to the leg modules. The steps in the process are linear and no complicated multiple tasks running. For sure the each step gets the correct timing, a simple while loop time delay function is implemented. When a step requires a long time delay, all other controls are blocked. So it pulses during each moving step by querying the ultrasound signal.
Although the PIC32 plays a simple role in the controlling the FIGARO, the entire system motor control is complicated. Each leg has 3 servo motors (18 of them totally for six legs) and they have very fine control. However, the l8 leg servo motors are not directly controlled by the PIC32. It must have another smart component (microcontroller or other smart device) on each leg for getting the commands from the serial port, interpreting it and controlling the servo motor action. The FIGARO reallocates those control complicities to the peripheral modules. Now the question comes. Although the FIGARO Bill of Materials claims only $57.81 in project parts, the actual cost is much higher than this figure. If each leg servo module costs $50 (could be much high than this), the entire system components cost can be easily over $1,000. I am not sure if buying the high cost components to build a competition project is fair to the others.
I checked the contest rule of using additional components and it says: “Components used in the design need not be purchased from Digi-Key, however all such components must be ones listed in the Digi- Key online catalog at its website at www.digikey.com at some point during the period of the Contest. The total value of the components in the bill of materials prepared by each Contestant must not exceed $400.00 in value*, as calculated using the 1-piece pricing for the components published on the Digi-Key website as of the date the bill of materials is submitted by such Contestant to Sponsors.”
There is an amendment dated on 5/9/2008: “Although ‘hardware components’ means electrical and electronic components usually assembled on a Printed Circuit, Wire Wrap or Prototyping board, the $400 BOM cost limit does not include this board or any cabinet. Nothing in the PIC32 Starter Kit or PIC32 USB Starter Board, I/O Expansion Board or Prototyping board counts towards the BOM cost limit except for any hardware components on those boards that the Contestant uses in his or her design.”
According to the descriptions in the contest rule, the FIGARO servo modules and markers should be in the BOM and is subject to the $400 limitation. If that is true, the FIGARO qualification for the competition is a big question mark.

The servo motor control may be done by a device like this:
http://www.lynxmotion.com/Product.aspx?productID=395
I used one of these in a 4-leg project a few years ago. It has an RS-232 serial interface, and will control up to 32 servos.

I think it could be argued that the project is "a controller for a robotic platform", and therefore the platform itself is not part of the $400 BOM. But it depends on how he described the project.
When I built my robot project, I think the servo motors cost $20-25 each.

My biggest criticism of the Figaro project is that similar robotics projects have been done 100's of times. The interesting logic for 'walking' and such is well-documented on the web. Several companies sell kits for 'hexapod' robots.

Pete

I have seen so many different hexapods out there that I had scrapped my own design (which was based on a Pic16F877, and actually controlled all 12 servos!!), just for the simple fact that it has been done too much. Now if it was a Biped robot who can swing a weed whacker to cut designs on the lawn (like the ASIMO robot), that would have been cool, or if the Pic32 actually did the controlling and not some aftermarket servo controller.... maybe if the blade was sharper, or spun with higher RPM's maybe it could cut the grass in one pass.

B. Morse

Hi,

Thanks for your support.
I have a lot to learn about robot motion from your project.

It would have been really great if I could meet you to discuss more on this but I have crossed my fingers for now.

Best of Luck,

Warm Regards,
Boseji