Let's Make Robots!

OddBot's Quad Bot (a Spider with 4 legs)

Used to demonstrate to students how easily a robot can be built.
Quadruped2.zip1.19 KB

This is a very simple quad bot I built to demonstrate to students how easily you can built a walking robot. DAGU is planning to hold a class on robotics for young students at the GuangZhou Science Center. Using body segments from my caterpillar robot I made some simple robot legs.

The caterpillar segments are great as they have mounting positions for the servos and 3mm LEDs plus screw holes that you can thread wires or small cable ties through.

The first video shows the robot using a gait for rough surfaces. The robot shifts it's weight from side to side and balances on 3 legs while it raises a 4th leg to move forward.

The second video show the robot using a much faster gait for smooth surfaces. In this case diagonally opposite legs are raise and moved either forward or backward depending on the direction. As the robot is trying to balance on only 2 legs one of the raised feet may still drag on the ground which is why this is only suitable for smooth surfaces. I have attached the code used in this video.

Click on the photos for full sized images:

For this project I am using a prototype of my Spider robot controller. This is basically a clone of an Arduino Mega designed specifically for running a large number of servos and sensors. As such it includes a 3A switchmode power supply and all 70 I/O pins are terminated with a 3pin, servo compatible male header as well as a female header for jumper wires. The spacing of the female headers allows custom shields to be made from a standard prototype PCB.

To drive the LEDs on the legs I could have connected them to the controller. This would have the advantage of being able to program different patterns and adjust the speed but in this case the LEDs are mainly for decoration and to ensure no one accidently stands on the robot. For this reason I decided to use a very simple LED chaser circuit.This simplifies my code.

This is surplus stock from a toy factory which I have soldered a male header pin so it can be used with a breadboard.With the header pin soldered on it was very easy to connect to the Spider controller for power using some female to female jumper wires. I used a cable tie to hold it in place but the circuit is so small a bit of double sided tape or hot glue would have worked fine.

The hardest part now is developing a good walking motion. The videos show my first attempt where the robot balances on three legs while it lifts a fourth. Although this worked it was a non-symetrical gait that causes the robot to drift to one side. I am now working on a new, symetrical gait and in the near future I want to add pressure sensors to the feet so I can teach it to climb over rough ground.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

1. Is the switch in the ON position? The board will work from USB power but the servos will only move weakly.

2. Are all your servos plugged in correctly with the ground pin being on the outer edge of the PCB and the signal wire closest to the processor?

3. Have you got at least 7.2V? Use a multimeter!  If your voltage is less than that then your batteries need recharging.

4. Post a good quality photo of your setup so I can see how you have wired everything.

1. The switch is on, but the servos move weakly.

2. Yea servos are plugged in right because the bot moves nicely when using power from USB.

3. I tried 7.5V (5xAA 1.5V) with the same result, I am yet to verify it using a multimeter(because I dont own one).

Programs which involve only LEDs work perfectly via batteries but servos act like they got polio.

4. Here are some pics:



the outputs with just 2 wires are LEDs rest are servos

the outputs with only 2 wires are the LEDs rest are servos


Didn't notice this one is on production until today I saw it in robotshop! Congrates!

Thanks HM. I can't take too much credit, the caterpillar parts used for the legs were designed almost 2 years ago, the laser cut chassis was designed by our manager for the Chinese market.

I just slapped it all together.


That was fast walking quadbot. I am sure to get one or two of this board cuz my 2560 doesn't work like your tips. Perhaps it might damaged or? Anyway, please keep us updated when spider is out.

Wow, what an improvement!

the action in the 2nd video is a vast improvement over the 1st video.   that little guy moves so smoothly, is it truly creepy (in a good way :-) )

is that just from improving the code, or did you make hardware changes also?

very impressive.    this sets the standard for all quadbots to be measured against.

Yes the new gait is based on a lizard running. It is purely software changes. This gait is not suitable for rough ground as the robot cannot balance on two legs very well without using a gyro sensor and more servos.

The first gait was the same used by little dog and is better for rough ground.

for rough ground, can the lizard gait raise the legs a little higher?


The issue is balance, not how high the legs are raised.