Even a bot that just traces the track shape and size and beeps every 5 metres would be cool. Walk behind dropping a cone at each beep location.
1 Like
oh man. now i want a quad-copter cone dropinā bot.
2 Likes
keep injecting your pragmatism. itās good. (tho itāll never get my my quad copter bot)
2 Likes
Iām not convinced that weād need RTK. The huge advantage we have is that we control the entire space. We can put up base stations anywhere we want and then just use triangulation.
I think you could do it with sweeping lasers and PPS. Iām not really good at maths, but if you have 4 lasers, one at each corner, that spun at some known RPM not only could the base stations figure out their location automatically but then any robot operating within the stationās box could also know its location.
somewhere between the PPS signal and the time and angle of detecting the lasers (maybe each one operates at a different wavelength or maybe they alternate sweeps) and some sort of RF component (ESPNOW) so they can talk to each other.
So the robot knows where the center of the box is, it can then automatically build the track by fitting it within the bounds of the box. Or a human could do that step and then just push the go button.
We just need a clever way to determine the angle of the laser strike. Or maybe not. If the base station transmitted itās angle at T=0 and the spin RPM was known then you should be able to determine the angle between you and that station. If you can do that for 4 stations thatās math even I could do.
Frickinā lasers, ESP32ās and GPS modules with PPS are cheap.
The best way to collect the cones would be to build cone scoopers that you mount to the front of your esk8 and turn the team loose at the end of the event.
3 Likes
What about augmented reality? AR over a live camera feed should mean that even if it has unexpected deviations it could accurately place the cones.
2 Likes
Thats what I was thinking, having the app first to āhuman botā might be a great start 
Actually if the app uses external GPS like Racechrono it might be pretty accurate.
The app just needs to show you if you are close enough to the zone where to place the cone. For example with starting/selected point of the track layout being at your current location.
4 Likes
I think it is 5x less convenient than charging and setting up a cone bot but 50x easier to implement.
It looks like this.
What Tuckjohn said about parking spaces is genius since you canāt really compare where you are to the lines enough but you can compare the lines to the terrain.
edit: oops forgot the picture
1 Like
Also forgot to mention VESC has UAVCAN baked in, you can control VESC controller(s) via Ardupilot.
1 Like
Maybe itās been mentioned but what about a bot that could drag a chunk of chalk? Then humans thrown down the cones on the line.
1 Like
Bringing this thread backā¦ really sorry for delay just life gets in the way
I had an idea of manually making pin coordinates on google maps for precise localization and then stringing those coordinated pins together for a bot to move to
Super basic but some sort of manual path needs to be made for the robot to follow
1 Like
Manually making a trackmap also seems reasonable for anyone else to do, the annoying part is just spending the time hashing that out, so its really up to user to make the most of it
but if there are people local to you obviously share and create a repository of track layouts
I will be happy to have a place to show that kind of stuff in this thread once I get to that point
I will share my progress as I get deeper i just wanted to resurface this thread so that you guys know I didnāt give up on it yet. I am especially motivated to bring this to life because of my spine injury i do not want to bend down for cones like ever again
Coming from someone also with spinal issuesā¦ snag one of these. Not perfect every time but as long as the cones arenāt super jacked from years of use it does a pretty good job of picking them up. Kills me I waited till after throwing almost an 100 intro sessions before I bought one 
.
Another guy bought the cheaper one that only has a toggle bolt type head and it doesnāt work as well so Iād say spend the extra $10 for this one.
1 Like
Iām happy to see this thread come back to life.
It looks like I never got back to you guys on the location modules. They are DWM3000ās. They are expensive ($25) but if you use them just for corner markers and put one on the cone drop-bot then itās only an up front cost.
I did some work on this for a different project but only got as far as setting up the markers and doing a proof of location. However, with Claude Code at our side (no, really, a valid use for AI 
) Iām pretty sure that we could have a system up and running that consists of 4 corner markers and a cone drop-bot that follows a list of coordinates relative to the markers.
Iām sure itās doable to drop them. Picking them up is a different problem. Probably a visual problem since the cones are all the same color. Not that hard but not a territory Iām familiar with. I think the hardest problem would be picking the cones up if some of them are tipped over, etc.
Also, now that I think about it, it might be a challenge to get it to drop them without running over existing cones. sort of a variation on the Traveling Salesman problem
IMO the placing of the cones in the exact same position each session is the problem to solve.
We can pick up 200-300 cones in about 2 minutes when we have at least 4 or 5 people doing it honestly. And when weāve had big groups of 15 plus is hilarious how fast we pick them up in comparison to putting them out.
Iād like to see a bot that can make a small chalk mark every couple cones along a pre-designed GPS track. That way we could send it out while we clean the lot with push sweepers and leaf blowers.
Then after cleaning we would have little marks to guide us in placing cones.
Not against a more advanced bot that can throw cones, pick up cones, make cookies etc but that seems like a big set of problems to solve in comparison to a chalk bot.
At some point we got to consider making the cones themselves robots. They would only really need two motors.And with the price of an esp32 and DC motors it could be feasible.
Id use an optical flow sensor to track position of each robocone
There are obvious downsides that can be engineered away.
You know itās funny that you should mention that, I was thinking that out of an outcropping of this one could create a Bot that did gigantic chalk drawings in parking lots. Nothing stopping us from doing a color 60ā image except, you know, chalk supply.
thereās like an 85% correlation between these 2 projects.
</scope creep>
One thing we need to think about though is standardizing the way in which the cone positions are documented. My first take is that youād have 0,0 as the exact center of the field and and an indicator of which way is North, then a spreadsheet that just had the XY coordinate of each cone relative to that center oriented to North.
Doesnāt matter what format we use, so long as we all use the same one.
do you think an optical flow sensor would work on blacktop and concrete?
I think keeping all the batteries charged would be a challenge. But the great thing about being esk8ers is that we have tons and tons of old 18650s around.
1 Like
If you search around on YouTube you can find chalk bot type projects. Theres even an off the shelf product for painting soccer fields too.
Just need someone smart enough and passionate enough to build something.
IMO making small chalk marks would be preferable since a majority of us are just barging these lots.
We try and have a leave no trace approach
1 Like
One of our local Auspev organisers Chloe has an app that plots the standard tracks on your map for you.l using racebox and dragy. Check out
Electric racing life.. I donāt have an apple so canāt link
Chicago looks like a bit of a bitch thoā
Iām happy to help out as time permits (of which I donāt have much currently) but itās @CHADFARTHOUSEās project so Iāll wait for guidance.