This project aims to build an "smart" robot lawnmower using open technology. Follow the build log here : https://www.robotshop.com/forum/my-first-robot-t16856/60
Features -differential drive using two sensored brushless (BLDC) motors -driven by two open source speed controllers (VESC) -uses Field Oriented Control mode to have near silent operation -no perimeter wire (main goal) -3D-mapping with camera and lidar -precise location using RTK -9DOF (9-axis motion tracking module) -regenerative braking -traction control -solar cells (only as a extra bonus power supply) -path planning, different cutting styles (diagonal, vertical and more) -self-adjusting cutting height
A robotic lawn mower is an autonomous robot used to cut lawn grass. A typical robotic lawn mower requires the user to set up a perimeter wire around the lawn that defines the area to be mowed. The robot uses this wire to locate the boundary of the area to be trimmed and in some cases to locate a recharging dock. So this project was born in my mind when I bought my first house and invested in a Robomow RC312 for the garden. Running it for almost 2-years I encountered lot of problems with the existing “brains” and basic functionality of the “robot”. So I started thinking of ways to improve this. After a some research. It seemed that technology existed to build a really smart robot that could make its own decision and adapt itself to the garden. Looking up different solutions on how to solve this, I found the RTK technology was very interesting. But a bottle neck existed, the GPS signal needed to be quite strong to receive centimeter accuracy (would not work well with trees and other items that cover the satellite reception).
Combining the RTK with Lidar sensors, Intel® RealSense™ 3D Camera and the odometry coming from high-resolution rotary position sensors on the motors would maybe give the needed accuracy for autonomous lawnmower that could 3D map the garden and make its own decision based on the environment with centimeter accuracy.
The goal is to have the ability to create virtual boundaries around objects and the flexibility of adapting to environment changes without today's hazzle with perimeter wires.
I believe that in the next few years all new mowers will have the ability to navigate without the perimeter wire. A transition to actual autonomous robots will happen and this is my development project entering the world of robotics.
There are quite a few mower projects to learn from, but based on the way you describe the project, it’s a lot less about mowing and a lot about 3D mapping and computer vision. If you have the skills / expertise to do this, that really sounds like an amazing project, and creating it as a lawn mower almost seems secondary. The sweep can create a 2D map, but you’d need to upgrade it to the optional 3D printed variation. If you create it as a “module” which can be added to other robots, you have something with commercial potential.
I want to build a planetary rover that zips around the universe taking selfies. I have two Jetson TX2 development kits – TWO! Not to mention a wild assortment of motors and extraneous robot parts. But what good does it serve the community if I can’t channel my thoughts into a working demonstrable prototype?
Here’s a wild idea – segment your project into sub-assemblies: connect the Jetson TX2 development board to a battery source. Label it, oh I don’t know, something meaningful like – MCU sub-assembly. Post the completed work to the gallery and voila! You have a working part. Post the next assembly. Build up the assemblies and eventually you arrive at robot lawn mower.
In short – big deas make great fodder for discussion board.
The gallery is best reserved for completed works or any parts thereof.
As you said, there are still a number of advantages to using perimeter wire, but you certainly look like you have the skill and passion to develop a solution which does not need it.
Thanks CB. Since this is my Thanks CB. Since this is my first robot project, it’s a lot of stuff to learn, but eventually I think I will get there. I am touching many new subjects which are far away from my current knowledge, but that’s also why I think it’s so interesting and enjoyable.
Hi, I too am interested in robotic mower solutions. I have thought about using radio telemetry from three transmission points to develop an accurate positioning system, but I have not had the time to do the work. In my area, I believe the biggest problem is going to be theft. If the robot is working autonomously 24/7, it will have to be able to protect itself from being stolen. I don’t have a solution yet… light weight, easy to steal, but quick. efficient, and nimble… Heavy(600#) then it would be hard to steal, but costly to run, big and bulky, damaging to the yard, and more dangerous, also more costly to make. I am coming to the conclusion that the bot needs to be simple with low cost parts so if (and when) it gets stolen, another can just take it’s place. Keep the expensive processing in the transmission towers, with the mower being a mindless drone that merely follows instructions and moves where it is told to go. Battery, one motor powering the cutting disc flywheel and drive system, two drive wheels with on/off clutching, proximity sensors, radio transmitter/receiver, and inexpensive Arduino pro mini with three relays.
