Seeker II Autonomous Mini Sumo

Seeker II is my next-generation stealth mini-sumo. It is under 10 cm x 10 cm in its normal start position. It is 1.5" high, and weighs in at almost exactly 500 grams. And it is fast, very very fast :-)

Specifications:

Here's a small movie (119 KB) of Seeker II trashing Drone. This was before I had the shell done, and when Seeker II weighed in 200 grams lighter than Drone.

The Yahoo group Robot-SUMO-Bash was set up for the mini-sumo competition held in Burlington, Ontario on December 14, 2002. Seeker II had its competition debut at that event, and it won first prize.

JonHy-Bots-wins-1.wmv (3.16 MB) - a clip containing all the bouts Seeker II had, along with a few others. The final battle was against Marauder, my brother's mini-sumo, which was the 2002 Canadian champion mini-sumo -- Marauder won both the Western Canadian Robot Games in May and the Eastern Canadian Robot Games in October.

BuildFest at ToroBotics 1-WinMedia.wmv (4.4 MB) - this is a clip showing some stuff that happened at the Toronto Robotics Group BuildFest in December 2002, a week before the competition, including some footage of Seeker II practicing against a few other mini-sumos there.

March 23, 2003 - Seeker II won first prize in the Light Autonomous class (class "F") at the OCAD sumo tournament on March 22 in Toronto.

April 21, 2003 - I decided to release the source code for Seeker II. You can find it here... This code is written for the CCS C compiler

May 8, 2003 - I've had a request for more detailed information on the electronics, so I'm including a block diagram and a schematic, both in GIF format.

May 27, 2003 - I've got a few videos from playing around after the Western Canadian Robot Games (I got knocked out in the finals)

Seeker II won honourable mention in the Circuit Cellar Mad Dash for Flash Cash competition held in 2002...

ECRG/CNRG (2005-2007)

My friend Jerome (who won second place) filmed the final match between Seeker 2 and Slice, his robot...

Finally, some pictures (click for larger versions):

November 2007 - Here's a shot showing all the pieces, mostly taken apart. The left motor is taken out in this shot, and you can see the delrin sleeve I machined to hold it in place inside the motor mount tube. The 6-cell AAA NiMh battery pack is also visible in this picture - I took it apart, and re-formed the pack in a triangular shape so it would fit. All the battery and motor wires were replaced a couple weeks ago with new 20 gauge silicone wire from BaneBots. The front bar is visible here as well. It holds the range-finder sensors, the close IR sensor, the power switch, the edge sensors, and the circuit board.
November 2005 - Here's a shot showing some internal details, including the circuit board, and the inside of one of the wheels. The 2 nine-volt batteries have been replaced with a single 6-cell AAA NiMh battery pack, and it works much better.

Also in this picture, barely visible, is my new front center close-IR sensor, or at least the wires leading from it to the circuit board. You can see more detail of it in the picture below.

Here's a front view, with the cool graphics, and the new front close-IR sensor. It uses a Sharp IS471F sensor, which has a really nice feature - it generates a 38 KHz signal to drive your IR LED with.

The basic concept behind the close IR sensor is the robot will only ignore the white edge of the ring when it can sense something with the close-IR sensor.

This is it -- The shell is on, painted, and ready to roll. It sure looks a lot nicer with the shell on and painted !!

Now I need to fix a few bugs in the code, make the search algorithm a little more aggressive, and I'll be ready for the 2003 Western Canadian Robot Games...

Update March 23 - The code is done, the bot is very aggressive, and I look forward to seeing some real competition in May...

Here's the back view... you can see the big, fat, grippy tires. The start button is at the top, and there are a pair of LEDs at the very back -- the green one is the power indicator LED, and the red one is a heartebeat indicator. At the bottom in the middle is the serial tether plug.
...and the front view, low and sleek. You can see the "eyes" -- a pair of GP2D12 analog range-finder modules from Sharp. I removed the standard connector from each one, and soldered a new standard 0.1" connector on, which makes it smaller, and much simpler to remove and replace when I take the whole thing apart.
A side view. The holes in the side panels are for the battery clips to bulge out -- otherwise it would end up being wider than 10cm. The batteries I use are standard 9 volt NiMH batteries, with the metal case stripped off and the bottom pad removed. This reduces the length by about 1/16" for each battery, which allows me to put them end to end and still fit inside the 10cm width.
Here's a picture of my new close IR board, mounted between the Sharp sensors on the front. It has an IR LED, a Panasonic IR detector, and a normal LED to help me tune it. The schematic comes from my brother's robot Marauder, although I have managed to fit it in a much smaller space.
Here's a top view of the board, attached to the front bar. The empty 8-pin SIP socket you can see at the top-left of the board is the connection point for the serial port and the LED/button cluster, both of which are in the motor mount.

Immediately to the left of the 8-pin socket is a 3-pin SIP socket, which will eventually house a transistor that will switch power to the sensors. The range-finders suck a lot of power, and there is no reason to have them powered up when I'm not actively reading them (like when I'm programming the chip, or in the debug menu, or waiting for the start button to be pressed). Currently I just put a jumper into this socket to provide power all the time.  

And here's a view of the bottom of the board. In general, I use red wires for positive power, black for ground, and white for signal lines. The only real exception to this are the wires going from the power switches to the circuit board connectors -- those are black/white, because those are the only color Radio Shack sells for two-wire combinations.

You can also see the 3-wire ribbon cables for the four sensors. There is a 1/4" space behind the big brass bar that is used for cables, the power switches, and the edge-detector sensors. The inch of space behind that is where the two 9-volt batteries go.

Here's the shell mounted before I painted everything. The top piece is brass, the side panels are aluminum, and the wedge is brass.
And finally, here's a picture before the paint job, with the top piece off. The new circuit board is lower in the front, and generally speaking is much better in terms of layout. I also moved the edge sensor resistors onto the circuit board, and added a big 1000 uF capacitor for the electronics.

The half-rod at the back, just behind the circuit board is a piece of 3/8" brass rod, machined to half thickness, and drilled and tapped for mounting.

Incidentally, there are a total of 5 nuts in this entire robot, and no glue -- all the mechanical connections are either friction fit or bolts into drilled and threaded holes.

If you have any questions or comments, please email me at Jon@huv.com.