Mechanical components for wheel steering

#1

Hi there,

First of all thanks for these great products you are selling. I like the quality very much.

In order to build more accurate and versatile wheeled robots it would be great if you could make the components necessary to build a steering drive.

The idea would be to have wheels that can be steered with a servo (in front or both in front and in the back). It will reduce the friction on the ground when turning (saving power and increasing drive precision) and it’s kind of a requirement for a 4wd non-square robots.

I am not a mechanical engineer so I don’t know what are the best solutions to achieve this but I think we need some kind of geared transmission and axis rotation control parts.

What do you think? Is there already a solution with existing components and a few tweaks?

Cheers.

#3

Hi nitro,

So glad that you like our products. Thanks for your suggestions, we will decide wheather we should add this part in our plan list.

Are you building a 4wd non-square robot now? Or what you are planning to build? Can you share you project to us? I think we can help to in the mechanical parts in some ways.

Thanks for your support! Cheers!

#4

Hello Eric,

Well I am currently experimenting with a 4-wheeled robot. I had initially backed your kickstarter campaign (deluxe without electronic) and I then purchased additional beams a few weeks ago. It allows me now to build a quite big structure (two layers using big beams).

I have also various electronic boards that I am embedding on it (arduino+raspberry pi, 2 cameras, various motion/location sensors). My goal is to build a fully autonomous vehicle capable of sensing the environment through stereo-vision. So far I am only considering indoor environment.

Challenging but very interesting… and rewarding when it works.

Currently I use it more as a research platform. I am slowly but surely building a comprehensive software library to interface all the hardware, learning in the process what works and what doesn’t :wink: I am actually quite new to robotics.

I see a few concrete applications such as watering plants, cleaning the ground, doing night watches to report intrusion… I think there are a vast amount of interesting possibilities. Though first I’d like to explore the basics.

So far I found a couple of problems that might be of interest to you:

  • There is currently no way to attach breakout boards / peripherals (such as cameras) to the beams. It would be very nice to have a few “fine grids” and a clever attaching system to pin down the extra boards. Probably different diameters must be supported for most common boards: M2, M2.5, M3, M4. Would be nice for example to have readily cut plexiglas of various sizes (with rounded corners) with standard holes for attaching to the beams and in which we can drill holes.

I solved this currently with wood. I cut thin boards to the size of my structure and drill tiny holes to attach the devices I want to install.

Just to clarify: I think the electronic modules you are making are great but I am more of the hacker-style. I want to understand everything down to each chip and being able to use any affordable and powerful technology available (ie. the Raspberry pi has an embedded graphic processor which will be useful to offload some of the heavy computations). For the record I am not even using Arduino library; I have written my own “advanced” stack on top of the low-level manufacturer api (asf). The only piece I didn’t write myself is the linker script :wink:

  • The motors have no feedback mechanism. Would be super nice to have a matching encoder by default (maybe outputting 360 pulses per wheel rotation).

So far I bought cheap light-based sensors and I plan to print b&w stripped disks that I will attach to each wheel. I could then fix the sensor right next to the disk and I should be able to measure the effective wheel rotation. But that’s hackish and a more robust solution would be welcome.

  • Proper steering drive. Skid steering has the advantage to be simple mechanically but it’s a pain to model mathematically. Initially I wanted to use control theory to predict accurately the drive of my robot but I didn’t manage yet to achieve a realistic model (working on any type of ground such as concrete, wood, plastic, grass, …).

I have no solution for that one yet. Hence my request. I think anyway that it would be more efficient to use a complete steering mechanism. It could be easier to install the motors on the moving part so no transmission would be required. Then a simple 90 degree rotation of the motor would suffice so that the motor is vertical on top of the wheel and take no space on the side of the robot. Of course the joint between the robot frame and the steerable wheel need some special hardware (servo + gears + axis). I don’t plan on putting extra heavy stuff on it (about 5 kg in total) but would be nice nonetheless if it’s robust.

  • Wireless power. As soon as you enter the autonomous world you want to let the robot charge itself. That’s not necessarily your core domain but that would be nice to have a generic solution.

I am currently experimenting with coils. That’s quite a specialized field though and I’ll probably need to read quite a few to get up to speed on that. If you could make a pair of systems (12V input - 9V output @ 1A) and a 8xAA NiMh battery holder with charging abilities that would be awesome. I know: much easier said than done…

Thank you for considering your customer needs, wishes and dreams :slight_smile: