Scholarship Robot Proposal: Garden Bot

Right now, my wife and I have a small garden(5'x7' if that). We have had good success at growing tomatoes so far. None of them have started to ripening so this would be a good test environment. The main idea would be to come up with a way to have the robot pick the tomatoes. One idea is to have a camera positioned on the tomato plants. The camera relays the information to a PC that does the visual processing. The PC sends general coordinates to the robot and the robot picks the fruit (tomatoes are really fruit). if the visual processing is not feasibility in the time frame available. The robot could water the plants at a certain time and a variable amount of water per plant. I have a couple of arduinos that can be used to aid in the functionality of the robot. If the arduino zigbee unit comes out it would be cool to use that as well.

sort by: active | newest | oldest
CryptLordGR10 years ago
Are the tomatoes fruits? I think they are vegetables :S .............................................. Very nice idea for the robot and difficult
neuromancer2701 (author)  CryptLordGR10 years ago
They are technically a berry. But back in the late 1800s the government was taxing vegetables and not fruit so they categorized it as a vegetable so the make some extra money. The grocers' definition is that a tomato is a vegetable based on the fact that fruits are sweet and vegetables are not. So who knows in the end, hahaha.
_soapy_10 years ago
Nightmare complexity! However, it is do-able, imo. However, actually harvesting the fruits like that is probably a bad idea. If you little robot decides that the tomato is ripe at midnight, and picks it, and you are rushed so don't check your tomato drop-box email until you get to work at 9am, then return from work at 6pm, you will find your fresh tomato is now rather manky, having sat in the sun all day. I'd go for a tomato tending robot first. Have watering done by a thin hose that is pinned up high, and can reach all over, or with a return cord system with the hose position monitored by software to avoid tangles. Have a system that firsts and removes slugs, and that can chase or otherwise scare off cats, birds, etc. as well as sending snaps of possible ripe tomatoes. This is easier to do, and is more likely to let you get to the stage of having a reliable robot that knows the garden boundaries, etc. onto which you can later add a tomato gripper and picker, should you decide the harvesting is still taking too long. Or you could ignore this because your motivation is to do something really hard from the off. Either way, it would be handy. I need to make one to flamethrower all the weeds as they try to grow in my gravel paths.
neuromancer2701 (author) 10 years ago
I have been looking into various ways to do the tomato identification process. Barring using another robot to scan the plants for the tomatoes. I am thinking of setting up a cheap web cam facing only one side of the plants (that is the biggest limitation of this idea). Photos could be taken and processed by a PC. I was thinking of using a Magenta filter which would identify the Red amidst all the green. Knowing that the camera is stationary. The geometry(IE a grid of coordinates overlaying the picture) of the tomato plants could be know before hand. The robot would receive the height and width location, it would have to figure out the depth. Ideally this could be used with a mobile bot, but since I don't have the time or money to pull that off this seems like a good compromise. The robot would still lack the intelligence to get the tomatoes deep in the plant. This would take a lot more processing power and a degree of manipulation that I can't address in this project.
neuromancer2701 (author) 10 years ago
Right now I am more concerned about the main functionality of the robot. The weather proofing won't be a big focus in the development stage. I think that the watering functionality will be my main focus. The idea of having to robots: one to water and one to pick is an interesting idea. The watering bot could also take video of the plants as it waters them. The video would be relayed to the PC for processing and the picking bot would be sent out as tomatoes are identified. I still think that finding and navigating to the tomato is going to be a big issue. I have a couple of ideas that address the picking function. Like you said there could be a soft but firm arm/appendage that would grab the tomato and I had the idea of twisting it to get the stem to break. The second ideas wound be some sort of jaws that would surround the tomato until it hits the stem then clamp down cutting it. One of my concerns is spacing and other tomatoes or branches getting in the way of the jaws. Thanks for your thoughts
Kira_Koenig10 years ago
This project can have great potential, but with major upgrades to the create. I thought about doing this for my grandmother's garden (slightly bigger than yours) and I got to thinking about all the upgrades to the gearboxes, motors, possibly the controllers (more power), sealing it up from moisture, adding tracks (so it won't bog in soft soil), a water tank (which would add massive weight to such a small bot), spray booms (optional), and possibly chemical resistance as well. It is a small robot and weight, battery life, and weather proofing are major factors. Even if it is a small garden. Picking the fruit I believe would add a lot of complexity to this project. I mean the robot would have to identify the fruit's stem, position it's arm to the stem, pull, and place the fruit in the basket. I know this can be done, it just maybe a little tricky to do. I mean you don't want the robot squishing all of your tomatoes. I am anxious to see how it goes, Good Luck!
neuromancer2701 (author)  Kira_Koenig10 years ago
I really wanted to apply this idea to a large venue. iRobot also makes the packbot and I think that could be used in a industrial situation. As for my situation, I plan on putting a small boardwalk around the plants to avoid the problems with soil and navigation. To avoid the bot carrying a lot of weight the docking station could refill the water tank. The biggest design issue for picking the fruit to me is actually finding the fruit or vegetable and getting reasonable close to it. If my problem is trying to find the stem I would be happy with that, haha. thanks for your thoughts.
Kira_Koenig10 years ago
-I see, I guess I was getting carried away. I didn't think to try an easy and simple solution like a board walk. I guess thats why we have forums. LOL -Although, what are you going to do about weather? It could rain at a moments notice (like it does where I live) are you going to weatherproof it or have it run for cover? -I did give the picking problem some thought after reading this comment. Could you use a robotic arm with maybe some foam padding that could get a firm grasp on the tomato, and still be gentle enough not to squish it?