 Hello everybody. So without further ado, we have Pierce Nichols here to talk about the laid-on project and This fine contraption over here. So please give them a warm welcome. Thank you Hi folks Jeremy and I have been and some other folks who are unable to make it have been working on building ocean-going sailing robots So our long-term goal is just build a Sailing robot capable of sailing around the world all on its own totally autonomously Since that's kind of a big a big lift in the shorter term We're working on completing the micro transit challenge and what that is is that's a transatlantic sailing race for autonomous boats Speaking of autonomous that has been held every year since 2010 No one is finished yet The maximum size of boat is eight feet, which means it's pretty cheap to To get out there and compete in it There's a boat in the water right now from a Norwegian outfit that I think probably has a fairly good chance of being the first to finish west to East We'll see how they do So prehistory on all this back in 2011 three Rick Gehansen who probably some of you know Prozed after after midnight at Haggerbot labs that wouldn't it be cool if we built the first autonomous boat to Travel all the way around the world No one done it then still nobody's done it today So a bunch of folks including Jeremy and some other folks who are not here worked on planning various parts of that but then we But that original effort didn't get much past the the paper boat stage Back in 2014 we decided that we really wanted to get something out for what was then the third tour camp out in Nia Bay and We went out and we bought the cheapest two-place fiberglass kayak we could find on Find on Craigslist and boy was it cheap and rightly so We brought it out to tour camp. It was not really ready for prime time and We tried launching it into the surf and it promptly rolled over so Well, that was a little depressing after that in 2015 We fixed the things that were obviously wrong. You know, we've made the battery attachment better. We improved the way the The way the hatch covers were attached Added external status lights basically everything that was annoying or really broke went poorly When we tried it at tour camp three we we fixed that We also made some changes to the software. So it was hopefully work a little bit better We brought it out to tour camp 2016 and while it was stable that time It didn't actually do very much 2016 2017 we got the software all fixed up Got it to navigate to waypoints got it to report its position and status up to the cloud we had some steering issues with it, but What I eventually put that that boat in moth balls was the fact that both the motor and the battery died I have a couple of maps of Where we actually went testing it around Around Elliott Bay right up side of Seattle the one with all the the blue arrows this guy is when we finally had it out like Navigating to waypoints on its own and all that the reason the reason all the paths are curved is because We had a problem in the code with calculating the magnetic Deviation so it followed curved paths to its target That was that was the trip our test on which we Killed the motor and the battery so We put we put that up on this We hang it from the roof of Hacker bot labs, and it's been there for probably a year and a half now In the middle of that We started building this sailing demonstrator, which is the boat that we actually have on stage here and this is pictures of it when we We started putting it together It's all laser-cut plywood It's the simplest possible design of a boat that you might actually be able to sail somewhere The the bow is plum the stern is plum the sides are plum, and they're just two circular arcs Nothing complicated The only internal structure other than some bulkheads to give it some stiffness is Places where the mast pivots go which is the you can sort of see the cases for them here And then the dagger board case here just the keel fits through and then you know Here's the hull after we got it all painted and fitted out and Other than scuffs in the paint it looks pretty much the same on stage here So I don't have any pictures of the first set of sails We tried to build for this because they really didn't work very well Well, not all that's better way to put it And you can see here what the wing sail looks like without its skin on it's got that kind of egg crate Structure that Jeff talked about that his boat is built with Again all laser-cut plywood and 3d printed stuff 3d printed fittings for the Food the tail booms and all that stuff And Then you can see there the completed one after we got it glued on coated with epoxy got all the sort of low points filled in with epoxy putty when we clamped it we just used it some Some straight pieces of wood with big foam pads under them across You know to sort of just clamp the skin onto it I had reason to cut into the wings you can see where I covered it over with silver tape recently to install some sensors and I discovered that the internal fillets which of course we did we'd done entirely blind right we just sort of gooped on some Epoxy with fillers and crossed our fingers that it actually all got into exactly where it's supposed to be and it's a Really strong structure So the two other really important parts of the boat are the keel Which is that big blue thing you can see down there our first cut of the keel was way undersized so this is so the boat you know Keel is not big enough the boat rolls over Which is of course what it did when we put it in the water that way So we made this larger keel. It's a steel tube filled with lead shot and it's got Again 3d printed end caps to give it a sort of nicer hydrodynamic shape The other beast here is the brain box for this So it has a beagle bone blue is sort of the main brain that has a compass and IMU on it, which is You know tells us which way the boat's pointing You know and more sort of very makery off-the-shelf stuff. We have a data fruit GPS there RC receiver so we have some manual override on it Cell modem so it can talk to the rest of the world and then a daisy AIS receiver so So we don't have to have any unpleasant conversations with a Coast Guard about getting too close to a ferry and this boat We've actually sailed in Green Lake so we've also dry sailed it on a on a pivot So you can see the Green Lake We have it on a nice tether The nice thing about the tether is that it meant that We couldn't sail it So we had to go swim after it or rent a boat to go after it We could just haul it back when we didn't quite have a quite have our sailing dialed in and the other one of these is attached to a pivot for Seattle Mini Maker Faire last year we got a big fan put it in front of a pivot put a random walk on the On the tails and it sort of sailed itself around It's a great visual demonstration of how the sailing works These tails that it has on it were the original 3d printed tails and they were too small So we replaced them with what you had seen out, which is You know, I keep seeing laser cut such and such this boat would not be possible without a laser cutter Some laser cut balsa with this absolute thinnest lightest fiberglass Surfacing veil that we could find Over it and even that, you know, it's barely there fiberglass made the the balsa wood so strong that you can't break it anymore Which is good because I'm always afraid that I'm gonna break these tails when I'm handling the sails So Let's talk about how the sailing works So as you can probably see these wing sails are free to rotate the Position of the sail relative to the wind is controlled by the tail so it works It's pretty much analogous to the horizontal stabilizer on an airplane in that it sets the angle of attack of the sail So if you And of course there are two of them and the reason that we have two of them is that it means we can delete the rudder Any anyone who's done any offshore sailing is probably familiar with rudder failures Also if look at one of the nice things about the microtransat challenge is that all the competitors are on a mailing list and they talk about why the boats didn't work and Probably more than any other single single class of failures other than getting tangled in fishing gear is breaking rudders So we designed that failure mode out And the way we designed it out is that we have the two wing sails so we can set the sails differentially right and If you set them at slightly different angles You get a torque around the center of the boat Right, which is you know just what a rudder gives you right an ability to apply a torque to the center torque to the boat And that lets you turn however you want to So this is our demo boat and Here's our microtransat configuration It's 2.4 meters long about eight feet It's made with a bunch of slotted together transverse and longitudinal bulkheads That make it very very strong and then the reason it is pink is because this is just router cut pink slabs of pink foam all glued together sanded to a nice fair surface and then glassed over and Then you can see our keel hanging down below which is again as you know astute just like this boat a big steel plate with a steel tube welded to the bottom and Filled up with lead shot. So the main computer for this is Beaglebone blue again because it's basically a can of robot, right? It's a Linux computer It's got Wi-Fi on board. It's got a mu on board. So we know where we're going and we can talk to the rest of the boat The wings have our Arduino mkr 1000s in them. That's again actually the same as this boat The nice thing about the mkr 1000 Versus another microcontroller with Wi-Fi like the ESP 32 is that it's significantly lower power which considering we're powering the whole thing with solar panels and In batteries is kind of a big deal I will say one more thing one more advantage of this configuration with the tails controlling the sails is that our actuator Power draws also is really low only a couple of watts per sail So talk a little bit about what sensors we have on it. So the Beaglebone blue as I said has an IMU So that gives us a compass direction and also it gives us the pitch in the roll of the boat as we're sailing along That'll let us among other things estimate sea state when the boats out in the middle of the ocean You know if it's do if it's doing a lot of this The weather is probably kind of rough It also has a temperature sensor Just nice for figuring out what the weather is barometric pressure, which also gives us some read on what the ambient weather is Beyond that in the main hall. We have of course GPS for location because Really, what else would we use? Everything else is much more complicated and expensive And also has a receiver for the automatic identification system and what this is is every large ship over I think the legal requirement is 300 tons Has to broadcast every minute or so its current location speed and course And that means that we can stay far away from fishing vessels We could stay away from cargo vessels when we're in the sound we can stay away from ferries You know not just because they're going pretty fast and they run us right over but because I think sailing an unmanned vessel near to one would would cause some uncomfortable discussions with the Coast Guard Then on each sail We have another IMU that measures the orientation of the sail and since we all know that We know the current tail command and the orientation of the sail we can estimate the wind Which is important to being able to sail the boat where we want it to go so Broken communications into two different categories There's the inshore communication, which is us testing in the sound and that's the an RC receiver Which is just the manual override You know we were we're pushing it off the dock. We want to sail it off the dock by hand We use that and then it has a cell modem so it can just hit the internet and say hey I'm from wherever I am Then when we get to offshore we're looking at rock block For an iridium satellite connection Nice thing about the rock block is it's relative It's cheap for satellite connection, which means it's about three hundred dollars worth of hardware cost and Something like a quarter of a kilobyte quarter per kilobyte for for bandwidth cost and we'll probably also put a spot tracker on board just as a backup GPS location Also an expensive hardware Okay, so I talk about how we're going to power this thing now My favorite part of the power system is that we're going to power the sails by molding in flexible solar panels I actually have one of the solar panels down at my camp down by the maker stage if you want to come see it I got it from a We bought it off a Chinese outfit called glory solar and it's like 18 watts and Like so so it's pretty efficient pretty efficient panel We have one on each side and I looked at it figured the easiest way to do this was to use each panel to charge its own string of batteries and Then we just choose the highest voltage Which gets us a little redundancy on the power system Main computer uses It's probably going to end up being too, but we originally planned for four Solar panels of exactly the same kind mounted on the deck same kind of You know one string per solar panel charging so now I'm going to talk about software for a while because Honestly any robot is a big pile of software with some appropriate hardware wrapped around it You know like bunny said that the hardware very much Governs what the software can do so you start with what the software needs to do and then wrap the hardware around it to figure out What that needs to know what that needs to be Now that's ideal right now the software is running way behind the hardware so it's built around something called moose IVP which is a It's kind of a middleware for robot for sailing or not sailing for maritime robots It's most of the development is out of the University of Oxford and MIT The real advantage of it for us is that it has a lot of modules off the shelf that do things like marine obstacle avoidance and marine collision regulations and You know how to do how to run a set of waypoints Things like that it plays well with the Beagle bone blue Which when I tried to install Ross on the Beagle bone blue it didn't work very well But moose you know first time it also has a lot of built-in mechanisms For getting the bandwidth of status reporting down Because again most of the people who are using this are on low bandwidth or expensive bandwidth kind of links Most of the custom software we're having to build for it is either configuration specific or Specific to our to our particular hardware The sail controllers no one's ported moose to Arduino yet, so we have a Arduino rest interface on them, which is very reasonably easy to use So so this is a software configuration, which is just sail it under manual control So green stuff is communications RC receiver We wrote a module that taught speak something called s bus, which is what the RC receiver speaks and then we have you know a mixer internally for that and then we have This module PA rest which goes and talks to the two sails which I put in red because they're actuators So that's the minimal Take it under RC control and go sail it you know go sail it around the bay right now to build up on that we want to add some GPS and AIS inputs, which is The Pgpsd what that does is that takes the output of GPSD, which is software package that parses GPS and parses AIS And it makes that available so we can log it P logger, which is here. It's paler because that's an off-the-shelf one Then can log all that stuff and then we also built an interface that lets it talk to It a fruit the ate a fruit IO service. I don't know any of you familiar with that Basically, it's a cloud-based, you know, hey write some data, right? And they have they have some nice dashboard tools. It's kind of limited, but it's good enough for what we're doing It's also relatively cheap and easy to use and it's It's kind of one more thing. We don't need to need to mess around with Nice thing about doing it this modular way is we could just keep larding on modules to do more and more so You know first time we want to take it out on its own We bring up helm IVP, which is sort of a It's a module that comes with moose and it lets you take a bunch of behaviors like say follow-way points and run them in in Inside the helm process and then it outputs based on on whatever you've told it, you know what behaviors you've told it to follow Outputs a go this direction this fast, right? Then to also make this work I have a little mux module that lets me switch between Manual and automated control things like that And then I'm currently working on a module called wing sailor Which actually turns go this direction this fast into set the foresail tail to this angle of the mizzen tail to this angle and do that thing but this stack of software is only really good for you know the Putts and around the sound right making sure the rest of the boat works Then when we want to actually go offshore We need to add a few more things We had need to add a module that talks to the rock block We need to bring up the contact manager module which is again a provided one which every time you see Hey, the AIS says there's a container ship. It's about you know 10 miles away. That's on course to run me over Contact manager spawns the avoid avoid Collision behavior, so it it sails away from the thing that's going to run it over And then we have of course a power manager module because we need to keep everything going So that's all the boat software we need some we need some shore side software to So Adafruit IO gives us sort of monitoring and control over cell modem capability I don't have to I don't have to mess with a server or anything like that I just sort of you know send them a few bucks every month and it just works Their data rate in their widget library a little limited though, so it's not great For future inshore Operations it would be nice to have a custom web app that it can go say here have some data and then it'll Cough the data back up And some sort of dashboard to be figured out when we're there For like for real offshore work The rock block, you know you send little datagrams and then the rock block Company server goes and hits a web app that you tell it to hit to dump the data out So we'd have to actually build that Okay, so Here's where we are on the boat right now We've got the hull all cat it up and we're working on getting all of our ducks in a row to get that All those pieces cut because unfortunately the largest pieces of wood for that are bigger than the CNC router We currently have access to The wings and the tails I've got the tails all made and glossed up. I just need to do yet more sanding On them to get them ready to go I've made all of the molds for the wings bars and the wingskins The cad for the wing is Somewhat variant from what actually from what I actually got built So I got to update the cad to reflect what I actually built On the electronic side. I've had the PCBs for and components for a few months and I've been busy with other things so that isn't isn't done I've written all of we've written all of the software except for the rock blocks and power manager modules Those are kind of future so we're working on getting all the other ones tested up and make sure they're all working Testing is naturally behind So now that I've yacked about software for a while I have pictures of building the various parts of the of the wings for the for the microtransat boat so the first thing I did was was the tails and I figured that the easiest way to to build the tails was to cast them I'm not sure. I feel the same way anymore But the first so the first part of making a mold is you make sort of a male plug and Again laser-cut plywood in this case with a polycarbonate Heat-bent polycarbonate skin over the outside of it makes a nice smooth mold And I was able to actually pour this rubber, you know two-part rubber mold around it That went pretty well So what I did with that was I wanted to cast in the the axles for the tail pieces so I Cast basically foam the foam cores of each each of the tails In the mold that you saw earlier And then I you know filled it up so they're pretty nice surface and then glass that over when I glassed it I got access to Vacuum bagging, which if you're doing fiberglass just like makes it so much better it's And what you do is you basically wrap it in a plastic bag and and suck all the air out so it clamps the Clamps the fiberglass across against whatever you're covering With fiberglass so you get a very not very dense It tends to suck out most of the air bubbles and It's very dense very nice Surface and it feels like it takes a little less work so for the wing we also needed to build a plug and Yeah, it's got you know a big aluminum piece of rectangle that sort of is the main strength for it and then plywood ribs to Form the shape now so the polycarbonate worked so well for the For the tail mold that decided we're gonna do the same thing for the For the wing mold Heating what was really about a four by eight sheet of polycarbonate up so that we could Bend it around the the nose radius We needed to get the to get some fire So we got a bigger steel tube the right size drilled a bunch of holes in it took it to a place called hazard factory Shove to propane burner down the whole length set it on fire and I have probably more pictures of us standing around waiting for it to get a hot enough up to forming temperature than any other stage Build because you know, we're standing around for 10 or 15 minutes usually I don't have barely any pictures of actually putting fiberglass on things because Wearing gloves to keep it off my skin epoxy it turns out will Will give you allergies if you have too much skin contact with it and then you don't get to use epoxy anymore because it's a sensitizer and You know if I'm wearing gloves covered in epoxy I'm sure so I'm not gonna touch my phone so I Mostly have pictures of it like this of putting the the skin molds or pulling the skin molds off of that Plug that you saw in the last slide So I have the leading edge Both you know while it was curing and afterwards you can see you the inside comes out as pretty nice finish And then you can you know vacuum bag a skin on to that under the inside of that and then it's just like the right shape and very light and then So the and here I have them one of the side the the mold for one of the trailing edges It was all laid up on it. Unfortunately. I didn't get any pictures of it after I popped it off But the inside looks similar to the other one Now there's a up the center of the sail. There's a there's a spar that's formed by two hat shapes and The way I made that one is actually the same size piece of aluminum is the wing but I used a router to cut a bunch of Foam you know foam sections, right? There's two inch thick foam sections and stacked them up glued them together covered them with one layer of fiberglass and then the first one I did by hand the second one I did I Applied that fiberglass with a vacuum bag the second one came out much much better Vacuum bagging is just awesome and then I hand laid a mold Here now this mold here Okay, it came off the plug. It was perfect. It was straight it was beautiful and I loaded it in my car before it was completely cured and as a result it got About it, you know, I'd say about a five degree bed a five degree bend 15 degree twist, which meant I had to do it again This is about five layers of fiberglass and I did it all in a single single shot By the time I was putting the top layer on the bottom layer the inside layer was starting to kick So it was uncomfortably hot to touch it Anyways, how am I doing on time? Well, I have lots of time for questions So team members, of course, Jeremy's here Alex couldn't be here Here Alessandro Is definitely involved with this but he's since moved to Colorado and Dylan Gray who also cannot be here and Let's see so acknowledgments of folks who've really helped us out one way or another Hacker bot labs where we're based The facility at Edmunds community college Where I've been doing a lot of machining and composites work It's maker space based out of community college. It's pretty awesome hazard factory and Rusty who probably some of you know Helped us with the with heat bending of things and welding other things Beagle board org has been kind enough to give us some hardware Leviathan and Frank Height have also been a big help to us and the Arduino folks have given us some hardware as well so links Blog that I should update a whole lot more often than I do Hacker bot labs also has a blog which I also need to update We have a github Github project laid on you can see the state of our software and Maybe even give us a hand on that. That's You know, that's one of our pain points is that need more software help If you want to just give us some money we have a patreon Under our old name hacker bot or hacker boat I mean and We're also on hackaday.io, which is yet another blog that I need to update Okay, so I Have about probably 10 or 15 minutes for questions. So Any questions? Okay, I think you went a orange jacket So I think the lead shot we ended up using was number four and That's just because it was on time The first size I saw when I was at Cabela's buying a bag and It's a lot easier to To use lead shot than it is to pour lead Right because I could just you know, it's a bag. I cap one end You know, I pour the lead shot in until there's enough and then I glue on the other cap I don't have to do all the setup for For you know doing a melt-and-pour with lead because it requires a fair amount of safety setup to do that without you know poisoning yourself and you had So we got the batteries got donated And replacing them with new batteries I mean, they're a big lead a big lead acid truck batteries So it's about 400 for the batteries and about a hundred dollars for a new motor So we decided we'd rather spend it on sailboats We may reverse that decision at some point And in the gray So just eyeballing it when we're test sailing it sailing it up at a green lake It looked like we were able to make progress up to about 15 to 20 degrees off the wind So it sails up wind really well because you can't left the sails And great It's really really easy to make Oh Really, that's that that is its advantage All of the shapes are very easy to understand because it's just two arcs It's it's vaguely sharpie like And looks a little bit like some of the advanced sharpies that Phil Bolger Designed, but it's even more simplified than that because it has no rocker on the bottom There's no curve on the bottom Yeah, it's basically just easy With this keel on it it though. What are lines about there? Yeah, you know an ocean this would be make a hook that would be really good at collecting Seaweed fishing gear everything else. So on the ocean going boat the keel is swept back and there's no protrusion So if it does, you know pick up some weed, it'll just go down and out rather than slowing the boat uh green hat in the back So we wouldn't we wouldn't want to do that right Because the the best position for the sail in the case So for instance, we were in a storm and we wanted to depower is to just Feather the tail so that'll point this directly into the wind Which is about the lowest air-resistant Lowest wind force that it could possibly have is when it's pointed directly into the wind so on this boat we I Can't think of any circumstances when locking it would be advantageous And for handling it's better to just take the sails out of boat Okay, you had a question then you So how do you kind of So we haven't done all the calculations for that So the expectation is that for for microtransat We can go several days Without any You know with with basically nothing The thing is is that for the North Atlantic if we did have that sort of extended period of No, no light we'd probably want to Feather the sails because it would probably be a big storm in which case if we feather up the sails we can shut down the the tail actuators and Really reduce our power draw and part of the you know what I was talking about with the power manager software that we haven't written yet It's sort of starting to shed off that all those loads. So we're just burning the minimum amount of electricity And ultimately the system Will be set up so that if we lose all power We just drift until we get some sunlight And you had another question Now I just For bad weather conditions is feather up the sails and let it drift Because ultimately because they're you know Ultimately an autonomous boat like this differs from a crude vessel in in a number of ways First of all, there are no unsealed deck openings, right? That's the first big thing right which means that if we have green water on deck we have green water on deck who cares, right? In those kind of conditions, we probably aren't getting any solar power anyways So we're and it's a sealed boat. So yeah, it happens The other thing is is that You know sailing vessels, you know like those You know age of sail Square rigors were not self-writing Right if they went, you know, if they healed too far over they were going down This boat will write for self-right from 180 degrees Right, so if a wave rolls it over a wave rolled it over. It's not a not really a big deal It helps also helps that it's small, which means the amount of You know the water on this part is going that way and the water on the other part is going that way it's a small lever arm and The difference isn't that the velocity difference can't can't be that large Just physically can't be that large On this boat probably not that much the axles on these are just eight millimeter stainless steel rods it wouldn't You could bend one of you if you really wrench on it you could probably bend one just grabbing it here and pulling real hard this boat is designed to hang around on Green Lake and verify that our software and electronics are sane Okay, I think you're you were first and then you Know there is no no connection So this boat since it's just a test boat. It's just battery power, right? so this piece of tape is covering the electronics compartment there and Actually, I'll take it off and Actually, what do you kind of come by after and then we can take a look at it So so we looked at different variations on that I mean it's definitely on our on our radar So we The problem with a lot of that is that So it's like you have a pendulum connected to a to a generator, right? It's one way to do that You end up with a fair amount of mechanical complexity And our design philosophy has very very much been design outweighs for it to fail Right, so things to their solid state. We don't have any slip rings. We don't have a rudder, right? so a Really a verse to adding anything else that might fail The other thing there is a simple version of that which is like a little strip of piezoelectric with a weight on it if you if you have a car built in like the last 10 years it has a tire pressure monitoring system and Those little sensors are powered by a little bit of piezoelectric stuff with a weight on it That that does this while the tire's spinning right and powers that And the problem is is that the free the roll frequency Isn't high enough to get an appreciable amount of power out of that So yes, we've totally looked into it and it's not really appropriate for our particular architecture and what we're doing and Gray you had a question does break well We do we have a we have a full IMU in each wing and We use that under normal conditions to sense the wind direction If the two sails suddenly start support, you know Reporting radically different values that clearly something has gone wrong The problem is at that point all it can say is hey, I'm broken Right, there's not anything we can actually do about it unless it's in Puget Sound then we go get it We will probably have a camera Yeah, well well, okay, unless someone gives us a whole lot of money We're not going to be live streaming anything for the middle of the Atlantic Yeah Yeah, hey if they want if they want to pay for that kind of bandwidth we will a hundred percent take their money, but Yeah, so cameras and hydrophones are two of the things that we have been thinking about it on there So for to me, that's you know, that's mission payload if it breaks it's loss of mission, but we still get the boat back the Everything shall be as simple as possible or is and we shall design out as many failure modes as possible is more You know ship systems rather than mission systems No, it's all on board But the thing is is that everything it needs to do it can do on board one of the software things We've been thinking about adding is wind chasing so because we can take publicly available wind data and National Weather Service publishes a global wind forecast for the next 48 hours every six hours and it's on a one mile grid on a One minute of arc grid, which is about a mile So the data to do that is definitely there However, I'm not quite confident enough in my own software engineering and test to to want to take the risk and looking at the wind condition, you know the average wind conditions across the Across the course. It's much better to just drift for a little while and then the next band of Storms or whatever comes through and you keep going so it takes a little longer than it would if you were actively trying to chase the wind But since just finishing is is the big challenge right now You know want to focus on focus on that So a big one is how it is rudder failure Another one is getting picked up by fishing boats to the point where I think it was last year the US Naval Academy named one of their boats trawler bait a Fair amount of them vanished without a trace, right, which is hard because you don't know what actually happened to it. I Is it a sail buoy, which is the Norwegian guys had a They had some kind of electrical failure last year and then the boat got found Which meant that they were able to diagnose it and try again this year They actually have a reasonable chance of making it For the West to East course and there they're actually out there right now if you go to the microtransat site you can You can watch their progress, which I certainly have been Okay Yeah, you So a lot of that is driven by what the actual contest rules are which is it's got to be safety orange pretty much We could put a put something on the sails, which is you know research vessel do not molest But a biggie, but I think a big part of our Of the way we're going to deal with it is we're going to choose the route so that we tend to stay away from like the grand banks And other areas that have particularly dense fishing boats The other nice thing whether nice thing is about a is is that vessels transmit what kind of ship they are So we can you know so we can have different rules for what to do about a fishing boat Versus a cargo vessel like if we're a kilometer away from you know an oil tanker or a Cater ship they're not going to stop and mess with us Being within sight of a fishing boat though That's bad for our health We probably want to give them you know more like 10 kilometers Basically teach the robot fear We So my feeling about having our own AIS transmitter is that's a lot of that's a lot of power budget and a lot of cost For what amounts to a steel me sign? so No so The North Atlantic is a big place And if you look at marine traffic comm you can look at like a continuous global AIS map Right and you'll see a couple of dots in the North Atlantic and they look like they're close together and you zoom in and they're 50 miles apart So for us, it's listen for a is Stay the hell out of everybody's way Yes Formally against the rules No, what it doesn't what it says is that you have to paint it For high visibility, right so people don't run into it And the other thing is is that if it breaks and it's like drifting I kind of like anyone who sees it to pick it up, right because if it broke I'd like to get it back you'd like to get it back so he could you know take it apart on the bench and see What the hell went wrong? Okay, any other questions am I doing on time? Okay, so I've been told that I need to wrap it up a little early so I can get the boat off the stage and Clear for the next person. So thank you all for coming. We'll probably be hanging out