Mecanum Lawnmower?
As I mentioned one of my projects this summer is to convert an electric lawnmower to a robot. Not only does this project appeal to me because of my laziness and not wanting to mow my own lawn but it allows for the creation of a platform which I can use for future developments. I started to dig into this project a bit over the past week and now have 2 different directions which I am considering.
1) A simple mechanical drive where I power the rear wheels and have casters on the front. This use of differential steering will be sufficient for my needs, but it doesn’t have that flare which makes a project stand out. On the other hand doing something simple (KISS FTW) may take the time required to finish the project more achievable.
2) My next thought is a mechanically more sophisticated system of using mecanum wheels on the robot. Mecanum wheels are ones that allow for holonomic movement(movement in any direction) You can read more about them here http://en.wikipedia.org/wiki/Mecanum_wheel One of issues with mecanum wheels is that they need to ensure contact with the ground at all time to work, which is why they are traditionally used in systems which roll on flat floors such as forktrucks. In the uneven ground of a yard or offroad the robot would require a suspension system, which is where the complexity begins. Normally suspensions are straight forward because their only requirement is to keep traction and keeping close to the ground is not a factor like it is when mowing the lawn. In the research I have done so far I have yet to find any off-road mecanum robots, in fact the closest I found is a concept for a mars surface vehicle. If I do go this route I plan on using the mecanum wheels from Andymark, as they offer 6inch, 8inch, and 10inch options with different modular hubs, which will make the challenge the implementation of the system not re-inventing the wheel.
This is the first paint style sketch of what I am thinking about if I decide to go down the mecanum path. Obviously the largest challenge is to keep the blade at a constant height so that I still have a good cut quality in addition to a autonomous robot cutting it.
Just curious, have you looked into the roomba modification scene? I’ve seen them modified for some interesting purposes, and I believe they even sell some sort of developer kit for them. It seems like the roomba has a pre-existing programming/circuitry to do exactly what you want, just meld the lawn mower chassis to the roomba brains?
I have thought about using a roomba style logic to run the system, but I think it is a bit more complicated then that due to the constraints of the system. One of the things which helps the roomba is that the floor is reasonably flat. It spirals out and every time it bumps something draws a virtual point which it uses to develop a full map. One of the constraints of this system is that the roomba’s wheels don’t slip, if they did it could skew the whole map. Additionally there are no physical walls on all sides to define the area like the roomba has.
In most of the robotic lawnmowers that are on the market they use a buried wire, just like an electric dog fence. While this does give the virtual walls it still has flaws. I have been considering using corner beacons, such as IR, wifi, or gps so that the robot knows it’s position relative to the world and wheel slippage does not have an effect. Each one of those sensor systems has its pros and cons, such as consumer gps is low accuracy (I am not going to do a differential system for this) IR is weather/temp dependent, and Wifi is a bit more complicated. Most likely it will be some hybrid system for navigation. The first thing I am going to focus on is the mechanical challenges of the system, as I can always integrate different controllers, sensors, and other electronics to run it.
Thanks for the ideas keep them coming!
Did you ever follow through with this project? I just ran across your blog and this article, while researching drive ideas for a similar project.
I was attempting to find off-road Mecanum wheels as an alternative to a synchronous drive system. The more I look at the problems you address, especially ground contact issues, the more a syncro drive looks like the way to go. In fact, it may end up being more mechanically efficient.
Any thoughts?
I never did this project because I got busy with some other things I was working on. I have done some work with mecanum wheels since then and really don’t like their inefficiency. There are quite a few other holonomic systems that would be better suited for what I was trying to do. While things like Swerve drives are more complex than implementing mecanums (if purchased) the contact with the ground would be easier to control and give you a much more precise movement without loosing 30% of your power instantly.
If you end up building something let me know, I would love to see it.
Greg