Does anyone know of any open source skid steering / differential drive platform? If not would anyone be interested in partnering up and developing one?
I’m mostly interested in having a platform to test different sensors, path following algorithms, software architectures etc.
There’s this: https://github.com/nasa-jpl/open-source-rover which is pretty detailed and actively maintained.
@XRobots and @cglusky have both done a bunch of OSHW - they may have something to point you too or get you started.
Simplest would be the Aion R1, which isn’t OSHW, but does run ArduPilot out of the box: https://store.aionrobotics.com/products/r1-ardupilot
I have made a small start on a skid steer rover, plan is to use mountain board wheels and belt drive. With a ODrive and brushless motors. Not had a chance to get much done yet. I have wheels and motors and have blocked out a rough shape in fusion 360. Plan is to make it from 3D printed bits and laser cut bits. Its going to be biggish probably 0.5m on a side.
Nowadays, ArduPilot should be compatible with every motor driver, so we can chose mostly any existing platform. Here are some example :
If you are in US, you can find some platform and good informations here : http://www.superdroidrobots.com/shop/default.aspx
You can also find some one https://www.robotshop.com/en/wheeled-development-platforms.html . Open source hardware will be hard. But with brand like Actobotics, you can have nice kit like https://www.robotshop.com/en/nomad-4wd-off-road-chassis-kit.html or buy parts to make you dream rover easily !
Thanks for all the answers so far!
@james_pattison thanks for the link to the rover. I’ve been eying the sawppy rover for quite some time too (which to my understanding is a simplified version of the NASA rover) that costs about 500 USD to build: https://github.com/Roger-random/Sawppy_Rover
R1 looks exactly like the thing I need! Being in Europe I expect it to be quite pricey when I add the shipment, and all the taxes to the base price but still it might be a good option.
@iampete thanks for sharing! I’ve been looking into ODrive at one point. Are you planning to use any wheel odometry information? Are you sharing the information about your build somewhere? I’d love to follow your journey!
@khancyr thanks for the links! Unfortunately I’m not in the US, so robotshop is the only acceptable option for me, however I can’t find a perfect platform there. I think you are right that I shold buy parts and make my dream robot! And then I can make it open source, because why not!
The plan is to add a proper driver for ODrive so we can take advantage of the encoders. I will probably do a build thread on here once I get a bit further on. Not had much time to work on it recently, If you interested we could collaborate but as I say I’ve not done much yet.
in Europe you will find plentys of distributor cheaper than robotshop for rover base, try to look in France or Germany (lextronic, generationrobot, gotronic, etc.)
@msadowski https://github.com/goodrobots/DevFrame will support a rover - you would need to come up with motor mounts for rover vs quad - since a skid steer rover and an H frame quad are essentially the same frame that is about it. i am way behind on just getting to v1 of the devframe quad so i dropped initial support for rover. biggest issue with using devframe for rover is it is not water or mud proof at all. good for inddor and/or fair weather outdoor use.
Sorry, but this will not work. Those long and thin arms will be to fragile to take the loads of a wheeled rover. Skid steering on any surface with friction will rip it apart. The wheelbase of a 4WD skid steering rover should be as short as possible, to make skid steering easier.
Sorry for the confusion. The DevFrame is just a simplified, more flexible version of BoxBotix, which is what is in the video. The DevFrame length is adjustable as are the length of the arms. So you can have whatever wheelbase and width you want. Or you can skip the tube on the arms altogether if you like, and bolt on typical bent metal mounts. Although I think you would be hard pressed to rip apart a carbon fiber tube with small hobby grade gear motors. We just designed 3D printed adapters to go from CF tube to aluminum tube that would just clear the gear motors. The outside end of the aluminum tube then gets 3D printed end caps that double as the motor mounts. Then it’s just a matter of choosing your shaft/wheel hub mount. We sourced our drive parts from ServoCity but there are many others selling similar stuff. Everything was either glued with 2 part polyurethane or bolted together using brass threaded inserts and machine screws as required.
The intent of the DevFrame is to serve as a starting point for hacking as the main frame is just a simple box truss put together with off the shelf tubes, screws, and 3D printed parts that can be printed from almost any desktop printer. It will not handle big loads as provided but it will scale. I use a larger, custom version for my personal commercial applications.
Edit: I get this question a bunch about the monster truck so will answer up front: The 12in wheels are CNC hotwire EPP foam. We used to do that in our shop before we sold it. If you can find 2in thick EPP foam sheet you could cut them on a scroll saw with patience. If I were to do it today I would design a 3D printed five spoke wheel (in multiple parts that gets glued/acetone welded together) with TPU/TPE printed bolt on treads. Why? EPP is expensive and hard to find, and, if you like classic cars (and maybe mullets), you know why a five spoke is so cool.