My first attempt at a hexapod walker.
@ May 2, 2010
- Actuators / output devices: 3 x 10.5g servo.
- Control method: Autonomous.
- CPU: PicAxe 08m.
- Power source: 3 x AAA batteries.
- Programming language: Picaxe basic
- Sensors / input devices: None as yet - Sharp IR ranging planned.
- Target environment: Indoor environment - Hard floor or short carpet.
I was so impressed by the video of the Pololu hexapod walker I thought I'd have a go at one of my own. This is a rebuild of my initial attempt using a PicAxe 08m. It's still held together with small dabs of hot-glue and will-power. I've refined the 'just make it do something' code and the new movement routine will be at the core of the program once I move this to an 18x. Currently it can dance and vibrate and can go forwards, backwards and rotate left and right at anything from an absolute crawl to full speed. Still a bit of code tweaking needed to get gradual turns left and right and optimise the leg sweep.
The component layout is different to the Pololu model, having the batteries underneath and the servos at the front. The beads on the feet give it a carpet capability but still some experimentation to do here. For the gait to work, there has to be some slip but not too much. (*Just discovered - the mid legs need grip - I've taken the beads off the middle and now it's brilliant on carpet*) The pics and video are pretty rough - short of time at the moment, but I will post something better as it develops.
I'll be using an 18x later as code space on the 'm' range is too tight for the things I've got planned. I want to include elements of my Instructables 'Piecax' project in here as well as obstacle avoidance with a Sharp IR sensor.
@ Sat, 2012-03-31 18:52
Sorry, no denounce on the
@ Sat, 2012-03-31 18:50
I like his rhythm.
@ Sat, 2012-03-31 19:32
Still a very much live project.
I did a lot more work on this guy, tidying things up and replacing the Picaxe 08M with a 20X2 and added an ultrasonic sensor. I had him quite cheerily strutting around the place and avoiding objects - very amusing to watch - but unfortunately I didn't document and post any of it here. Then he sat in a drawer for the best part of 2 years!
I rediscovered him again recently, replaced the batteries and was reminded just how cute this chap was. I've stripped him down to just the mechanics and intend to amalgamate the 2 control boards he ended up with into one, but my Rocket Studios Tiny Tank chassis kit just arrived today so I'm playing with that at the moment. (Hacking a Picaxe 28X2 board to do PWM - fiddly!)
@ Sat, 2012-03-31 18:50
I like his rhythm.
@ Mon, 2010-05-10 19:04
That's a great little hexapod!
It seems quite agile - and I especially like the colorful feet. Do you mind if I add your robot to the list at the end of my tutorial?
@ Mon, 2010-05-10 19:45
No probs.
I'm fine with that, after all, your tutorial was the inspiration to build this.
No work on it for a couple of days, but I'm building the controller and I'll be updating the pics and video soon.
@ Fri, 2010-05-07 18:19
Progress report - IR sensor probs.
I've gone off Sharp IR distance sensors. It works perfectly over the range it's specced for but appears that it is a characteristic of the device that it gives an in-range output when there is no object in range. Very annoying. There's at least one thread on this.
So I'm going ultrasonic with an SRF05 which I've used before and work over a larger range (but is a bit bigger). Still undecided whether to servo mount the sensor or just have it front-facing and use the legs to swing the body. Also, I've decided to go up to a 20x2 rather than an 18x 'cos it's only 2 more pins and a lot more powerful.
Work continues.
@ Fri, 2010-05-07 13:12
Cute
I like the feet. Mabe a soft silicon or rubber may work better.
@ Fri, 2010-05-07 14:08
Watch the slow bit.
I may try that on the middle pair but at the moment I've taken the beads off the middle legs. Look at the slow bit of the video - The geometry of the triangle formed by the 3 feet touching the ground changes as the legs move. There needs to be some slip to let this happen. I've found by experimentation that having the middle pair grippy and the other 4 sliding give the best result.
@ Thu, 2010-05-06 01:01
Quick Update
I've got the walking gait working really well now and it's happy strutting across carpet or a hard surface. Tonight I've been playing with the Sharp IR distance sensor and the next step is to build the actual controller with a PicAxe 18x (rather than the 08m I've been testing with) and get the rangefinder attached. Then comes the really good stuff . . .