Totally! The electronics supply world is crazy right now! We intended for this product to be much less expensive but that's just not possible. We're building the next batch right now and multiple parts are multiple times more expensive than they should be. One of them is $31 right now when it should be around $2. Ouch! This current price allows us to continue making them and having them in stock.
Hi there - nice board. I'd just like to mention that I've been working with I2C devices but have found they have various shortcomings and one being that lines are prone to picking up glitches that can cause the device to hang so it kills the entire bus until you cycle the power. The other problem is devices that connect to I2C operate in many different ways and is generally about getting and setting a combinations of registers which needs a custom software library to operate them. Also the address range is not that great if you want to run many peripherals. With small microcontrollers you do not want a copy of every library of every chip and besides you can't future-proof it. Anyway, what might be a better solution is something I was looking at which is part of the BLE specification. The sensor interface can be done on a one wire bus(called a micro LAN), the data-rate is low so it keeps interference down so you can have a longer bus, it has a larger address range and all the sensors fit a standard interface known as GATT. So lets say you want to pluck the value it currently reads then these will all be at a predefined addresses so one piece of software works for all. This system is designed for wireless bluetooth but on the market are many one-wire sensors. It's worth keeping an eye on this tech as it is developing fast.
hm… wouldn't it be better to just use a microcontroller for the realtime bits and then a pi for the heavier stuff? That's what Klipper (3d printer firmware) does, and it works very well.
Curious as to if this could be adapted to an RC aircraft with a tied in GPS to perform preplanned flight paths with camera support to be used as a semi-autonomous drone mapping and property management device for large farms and ranches. A pilot launches the plane then hands over control to the autopilot to map/record the property. Raspberry pi is a very efficient platform to base this on and 20 min mapping/observations should be well within the power levels of onboard batteries.
I can't see why that wouldn't be possible, since Navigator is already compatible with multiple ArduPilot firmwares, although it would presumably take some development work to get the mapping working. Maybe simplest to do live recording while operating, and do the processing and map creation in post. Depending on what's involved though it may also be possible to do real-time mapping with a light-weight computer vision pipeline, and/or use a camera with onboard processing that can do most of the grunt work (e.g. maybe an OpenCV AI Kit camera).
I am working with a missionary group. We have a need to modify the location stored in the return to home function to a new location in the flight computer I was wondering if you know, if this can be done with your drone control card.
Because your board does not have an independent microcontroller, I guess now the RPi and the Linux distro will be running of the control systems. How do you guarantee strict timing constraints for the controls (as it would normally happen with an RTOS)?
We currently don't use an RTOS patch in Linux. Some Linux autopilots do use the preempt patch that allows true real time performance. That may be something we consider in the future, but for now we have the flight control software set to the highest priority in the kernel scheduler and that is working really well.
@@BlueRobotics Thanks for the clarification. I do like the Navigator approach from a DIY/Maker point of view, because I could work on only one platform on top of Linux. Very convenient. However, for a more production-level application, I think having an independent uC would a more robust solution. IMHO, having a firmware with only safe-critical controls, like pos-hold and stabilization, and adding all the high level businness logic to a companion computer (like navigation, obstable avoidance) would be the sweet spot. Nevertheless, I do look forward to your new innovations. :)
@@ToshiLab Definitely see your opinion and it's valid! We do have to be careful with the software on the Raspberry Pi to keep things stable. Even with the Navigator it would be possible to keep the core tasks on the Navigator and use a separate Companion computer for tasks that don't make sense on the Pi. For instance, you could have a separate computer to do computer vision processing and share the results with the Navigator/Raspberry Pi.
Is the navigator able to attach to other sbc like lets say a ROCKPro64? It seems like such a cool product, but sadly the raspberries are all out of stock, with no end in sight.
Yes! The Navigator can control manipulators like the Newton Gripper with PWM signal, or more complex manipulators such as those from BluePrint Labs, which use a serial port signal.
Yes, BlueOS will run on the Raspberry Pi 3 with the Pixhawk as well. You can read more and update here: discuss.bluerobotics.com/t/blueos-official-release/12024
great idea... but $320 for just the expansion board :) keep at Blue robotics, make it cheaper, and for non-blue applications, and maybe you could become the new standard architecture for bigger fpvs... ?
Hi Brad, the camera is connected by USB or the CSI camera connector. Our BlueOS software, which is running on the Raspberry Pi, has a camera manager that detects the cameras and allows them to be set up to stream to a port. For instance, for video in QGroundControl, we stream the video to a UDP port. There's a lot of flexibility built in for different cameras and stream types.
The AX-12 servos use serial communication from a UART. There are three UART ports on the Navigator that could be used for this, but we don't think it's supported by ArduPilot, which is the default vehicle control software used on the Navigator. If you are able to write your own software, then it could work!
Hi HersheyFPV, yes it is compatible with CRFS. The "RC" input is a UART with RX only. If duplex communication is required, then one of the serial connectors can be used.
A bit overkill for what I am thinking about. Perhaps you can design a "Lite" model. Would have GPS and two motor outputs. One to control a trolling motor, and another to control a stepper or servo to control trolling motor direction by 180 degrees. If you have seen the i-pilots on modern trolling motors, it is obvious that a market exists for people to upgrade old $50 trolling motors instead of paying $1000 to $2000 for the modern trolling motors.
There are schematics, pinouts, and part numbers provided in the technical details, so perhaps there could be a community-led effort to design a smaller one for boards like the RPi Zero 2W (potentially without some of the peripherals, in a 'slimmed down' version as you've suggested). I'd guess that Blue Robotics won't be designing and selling another model in the near term given the current Navigator has just been released, and is sufficient for their existing vehicles.
Cool stuff but what I don't understand is this: I am running a 30.5 mm x 30.5 mm STMF4 FC (Mamba F405 MK2) which runs Ardupilot just the same as this however it is MUCH smaller and lets you solder your cables instead of having to plug them in with extremely proprietary JST-GH plugs. Sure, you need to solder on your sensor suite as needed however isn't this the more efficient way to do it? Never understood the size of the Pixhawk, either. I understand this is a development board but you're in the business of making these for small robotics applications, right? Live video from the Raspberry PI is sent over 802.11n/ax WiFi is extremely laggy, no? Better to use a purpose built unit like an analog 5.8GHz VTX or DJI's 5.8 GHz digital video transmitter that offer low latency at quite the range.
First off, the JST-GH is anything but proprietary, you can get sets of HUNDREDS of connectors for ~$20 on Amazon with 2 day delivery anywhere. It's nice having plugs because outside of the hobbyist space most people either don't know how to solder or are bad at it AND the plugs make it extremely straightforward and kinda foolproof (you don't have to worry about tx/rx pairing or sdl/scl pairing, it's done for you). And yeah, the navigator is larger but you have a TON more flexibility and processing power than you do on your F4 flight controller. On a miniquad it's fine having a slowish processor and one or two UARTs but that doesn't really work on a Rover/Sub. The huge number of PWM outputs, UARTs, and I2C is a must when you're getting into more specialized tasks (like the subs this board focuses on), and that extra flexibility takes space. In other words, your question is like asking why someone would use a dumptruck when a pickup and a shovel could do the same thing. Different tools for different jobs.
On the video front, it's perhaps worth noting that Blue Robotics works primarily with tethered ROVs, and the video for those vehicles is transmitted over ethernet via the tether (wifi doesn't work underwater). Other vehicle types are likely to either not have a live video feed (they may just do onboard recording instead, or have no video at all), or may have an external way of transmitting it (in the same way there's no RC receiver/transmitter or GPS receiver built in - that's unnecessary extra cost for some use-cases that can't make use of those features).
Hi Habiks, the price is high because the components are quite expensive right now due to shortages! We have to pay A LOT more for some of them and keeping the price high helps us keep it available. Hopefully we can make it more affordable eventually!
Chip shortage is a right pain for everyone at the moment, and if it was being sold at unsustainably low prices (compared to current component costs) it would likely end up selling out rapidly and becoming unpurchaseable / having extensive wait times like many other electronics at the moment (including alternative flight controllers). Hopefully chip shortage calms down soon so prices can reduce to friendlier levels, although unfortunately that likely won't be as soon as any of us would ideally like.
Guys, with this pricing policy you should teach Ilon how to make money. You really want $320 for a board with a couple of sensors? I wish people in my country would buy that for at least a third of the price...
Hi Mike, we totally understand your concern here and we didn't want to sell it for this price either, but with the current electronics supply chain situation it's our only option. We're paying 10x the normal price for some of the components on here and we'd rather have it available, even at a high price, then not available at all! Hopefully that changes in the future!
Hi Niels, are you saying there are VPN advertisements appearing on our website? We're aware there are some issues with technical details loading incorrectly at the moment, which we're working to fix, but there definitely shouldn't be any advertisements. If there are we would greatly appreciate any information you can provide about what you saw - please feel free to send us an email at support@bluerobotics.com
Hi Ognen, we're aware there are some issues right now with loading content correctly, which we're working hard to fix. If there's something else going on we would appreciate if you can send us an email about it at support@bluerobotics.com
Marine technology has traditionally been very expensive. Blue Robotics low prices have changed the industry. Not long ago this would have cost thousands.
@@GeoffCook1000 can't totally agree not totally disagree!. It is a very specialized area of course but, same company used to focus more on the hobbiest. I have a number of the older products back when they took open source seriously. Just going to have to look elsewhere it seems.
Do you know all hobbyists? Have you polled them all to see if most of them think this is too expensive? I don't think it's too expensive for what it does.
@@ZeeCaptainRon well I hope you buy one then. Give a try pricing out the BOM for the pi hat and then you might understand the markup here. Again I do value their contribution but it is overpriced.
@@SK-bl1lp lol yeah i have i was messing with ya but it is super reliable and a great upgrade from the pixhawk. The raspberry pi and the navigator are more organized and much more efficient. It was worth it for our underwater robot
We don't have a realtime kernel patch but we are running the autopilot code as the highest priority in the scheduler. There's a lot of power and cores in the Raspberry Pi 4 and this hasn't been an issue, but of course it is a valid concern.
@@BlueRobotics I really prefer having a two system approach. It doesn’t cost that much more to add another ARM MCU, and writing real-time state machines is easier than trying to outguess a scheduler. Bonus, each system can watchdog the other.
So great to see how far this company has come since the original kickstarter
I love everything Blue Robotics does. Ordering one of these now.
Looking forward to this when it gets a little cheaper 😀
Totally! The electronics supply world is crazy right now! We intended for this product to be much less expensive but that's just not possible. We're building the next batch right now and multiple parts are multiple times more expensive than they should be. One of them is $31 right now when it should be around $2. Ouch! This current price allows us to continue making them and having them in stock.
It's get cheaper if the board manufactured in china....lol
Cmiiw
That looks fantastic, well done! Most interesting drone thing I have seen in ages.
This is SUCH great news!! My friends and I wanted to build an ROV last year but halted due to costs and trying to find a Pixhawk
I wish it had a CANBus port or two.
Hi there - nice board.
I'd just like to mention that I've been working with I2C devices but have found they have various shortcomings and one being that lines are prone to picking up glitches that can cause the device to hang so it kills the entire bus until you cycle the power. The other problem is devices that connect to I2C operate in many different ways and is generally about getting and setting a combinations of registers which needs a custom software library to operate them. Also the address range is not that great if you want to run many peripherals. With small microcontrollers you do not want a copy of every library of every chip and besides you can't future-proof it.
Anyway, what might be a better solution is something I was looking at which is part of the BLE specification. The sensor interface can be done on a one wire bus(called a micro LAN), the data-rate is low so it keeps interference down so you can have a longer bus, it has a larger address range and all the sensors fit a standard interface known as GATT. So lets say you want to pluck the value it currently reads then these will all be at a predefined addresses so one piece of software works for all. This system is designed for wireless bluetooth but on the market are many one-wire sensors. It's worth keeping an eye on this tech as it is developing fast.
I had a Navio2 and it would regularly shutdown inflight and Emlid was never able to figure out what the causes were
I2C could be a problem. Every device needs its own library... :(
I love it!
Wallet just told me to jump;)
Awesome! Can't wait to get one
I would love to build my own drone one day
I would still need the raspberry pi for the camra feature
Absolutely amazing. Great job
hm… wouldn't it be better to just use a microcontroller for the realtime bits and then a pi for the heavier stuff? That's what Klipper (3d printer firmware) does, and it works very well.
Was also my thought.
Because the rpi4 could hang up.
A uC would not hang up so easily
Curious as to if this could be adapted to an RC aircraft with a tied in GPS to perform preplanned flight paths with camera support to be used as a semi-autonomous drone mapping and property management device for large farms and ranches. A pilot launches the plane then hands over control to the autopilot to map/record the property. Raspberry pi is a very efficient platform to base this on and 20 min mapping/observations should be well within the power levels of onboard batteries.
I can't see why that wouldn't be possible, since Navigator is already compatible with multiple ArduPilot firmwares, although it would presumably take some development work to get the mapping working. Maybe simplest to do live recording while operating, and do the processing and map creation in post.
Depending on what's involved though it may also be possible to do real-time mapping with a light-weight computer vision pipeline, and/or use a camera with onboard processing that can do most of the grunt work (e.g. maybe an OpenCV AI Kit camera).
This is awesome, I will have to figure a way to get one for my classroom!
I want to build a power wheel chair with this. Is that possible?
I am working with a missionary group. We have a need to modify the location stored in the return to home function to a new location in the flight computer I was wondering if you know, if this can be done with your drone control card.
Please make a small version for sub250g recreational builds. This looks awesome.
use b-roll for the close up shots.
Because your board does not have an independent microcontroller, I guess now the RPi and the Linux distro will be running of the control systems. How do you guarantee strict timing constraints for the controls (as it would normally happen with an RTOS)?
We currently don't use an RTOS patch in Linux. Some Linux autopilots do use the preempt patch that allows true real time performance. That may be something we consider in the future, but for now we have the flight control software set to the highest priority in the kernel scheduler and that is working really well.
@@BlueRobotics Thanks for the clarification. I do like the Navigator approach from a DIY/Maker point of view, because I could work on only one platform on top of Linux. Very convenient. However, for a more production-level application, I think having an independent uC would a more robust solution. IMHO, having a firmware with only safe-critical controls, like pos-hold and stabilization, and adding all the high level businness logic to a companion computer (like navigation, obstable avoidance) would be the sweet spot.
Nevertheless, I do look forward to your new innovations. :)
@@ToshiLab Definitely see your opinion and it's valid! We do have to be careful with the software on the Raspberry Pi to keep things stable. Even with the Navigator it would be possible to keep the core tasks on the Navigator and use a separate Companion computer for tasks that don't make sense on the Pi. For instance, you could have a separate computer to do computer vision processing and share the results with the Navigator/Raspberry Pi.
No CAN port?
No built in CAN port, but you could add CAN (or any peripheral) via a USB adapter.
I don't have any use for this but i am glad its here and i hope i will have a use one day, so i can get it. great project.
what's the speed of the control loop?
Indeed. Doesn't sacrifice any measurable performance compared with a standalone pixhawk?
Is the navigator able to attach to other sbc like lets say a ROCKPro64? It seems like such a cool product, but sadly the raspberries are all out of stock, with no end in sight.
can i used for AUV ?
Nice design
Would ROV manipulators go through this?
Yes! The Navigator can control manipulators like the Newton Gripper with PWM signal, or more complex manipulators such as those from BluePrint Labs, which use a serial port signal.
@@BlueRobotics Thanks.
I wonder how a gps yaw system would work. 🤔 no needful the magnometer
Very nice! But a I will wait for a while to buy it. The current price is insane!
Nice! Maybe someday when pPi4 are available again! Maybe Pi5s will be easier too get?
You can buy a Raspberry Pi 4 together with the Navigator 🙂
@@BlueRobotics I love what you guys do for the ROV world, but sorry, I am done with RPi's
Interesting product. Are u sure u want to name soft “BlueOS” when “BluOs” is in the market ?
Can blueOS be used with current pixhawk and raspberry pi configuration ?
Yes, BlueOS will run on the Raspberry Pi 3 with the Pixhawk as well. You can read more and update here: discuss.bluerobotics.com/t/blueos-official-release/12024
Wow this is impressive! il make sure to try it out
Maybe alongside an AI specialised board to make a perfect moving robot!
Noob here.
Is this better than cube orange?
Hello, can the Navigator card be used on a Rasberry Pi3 Model B+?
Hello, @hugolemiegre6384 the Navigator is an add on board for the Raspberry Pi 4. Unfortunately, it cannot be used with Rasberry Pi3 Model B+.
great idea... but $320 for just the expansion board :) keep at Blue robotics, make it cheaper, and for non-blue applications, and maybe you could become the new standard architecture for bigger fpvs... ?
Can you talk more about how the video streaming will be integrated?
Hi Brad, the camera is connected by USB or the CSI camera connector. Our BlueOS software, which is running on the Raspberry Pi, has a camera manager that detects the cameras and allows them to be set up to stream to a port. For instance, for video in QGroundControl, we stream the video to a UDP port. There's a lot of flexibility built in for different cameras and stream types.
@@BlueRobotics What is the stream latency? Any more info/documentation on this?
Please do you have a discord channel?
Does it compatible with ROS ?
Can this controls AX-12 servos?
The AX-12 servos use serial communication from a UART. There are three UART ports on the Navigator that could be used for this, but we don't think it's supported by ArduPilot, which is the default vehicle control software used on the Navigator. If you are able to write your own software, then it could work!
Hmm... Wonder how that'd work with a wingbourne drone.
does it work with jetson nano?
Hi! These do not with Jetson Nanos! It's something we are thinking about :)
It the RC input a full Uart compatible with CRFS?
Hi HersheyFPV, yes it is compatible with CRFS. The "RC" input is a UART with RX only. If duplex communication is required, then one of the serial connectors can be used.
If only raspberry pies were available... Looks great though!
You can opt to get a Pi 4 together with the Navigator :-)
I can supply, I have limited, be quick
A bit overkill for what I am thinking about. Perhaps you can design a "Lite" model. Would have GPS and two motor outputs. One to control a trolling motor, and another to control a stepper or servo to control trolling motor direction by 180 degrees. If you have seen the i-pilots on modern trolling motors, it is obvious that a market exists for people to upgrade old $50 trolling motors instead of paying $1000 to $2000 for the modern trolling motors.
Amazing!
Awesome!
That looks awesome. Can you make them a ton smaller?
Hopefully they can make a raspberry pi zero version, but I have a feeling it doesn't have enough IO
@@xWood4000 might be a slimmed down version. I've been wanting to make my FPV quad a little more fun with programmable maneuvers.
There are schematics, pinouts, and part numbers provided in the technical details, so perhaps there could be a community-led effort to design a smaller one for boards like the RPi Zero 2W (potentially without some of the peripherals, in a 'slimmed down' version as you've suggested).
I'd guess that Blue Robotics won't be designing and selling another model in the near term given the current Navigator has just been released, and is sufficient for their existing vehicles.
@@xWood4000 perhaps esp32?
px4?
Awesome
Cool stuff but what I don't understand is this: I am running a 30.5 mm x 30.5 mm STMF4 FC (Mamba F405 MK2) which runs Ardupilot just the same as this however it is MUCH smaller and lets you solder your cables instead of having to plug them in with extremely proprietary JST-GH plugs. Sure, you need to solder on your sensor suite as needed however isn't this the more efficient way to do it? Never understood the size of the Pixhawk, either. I understand this is a development board but you're in the business of making these for small robotics applications, right? Live video from the Raspberry PI is sent over 802.11n/ax WiFi is extremely laggy, no? Better to use a purpose built unit like an analog 5.8GHz VTX or DJI's 5.8 GHz digital video transmitter that offer low latency at quite the range.
First off, the JST-GH is anything but proprietary, you can get sets of HUNDREDS of connectors for ~$20 on Amazon with 2 day delivery anywhere. It's nice having plugs because outside of the hobbyist space most people either don't know how to solder or are bad at it AND the plugs make it extremely straightforward and kinda foolproof (you don't have to worry about tx/rx pairing or sdl/scl pairing, it's done for you). And yeah, the navigator is larger but you have a TON more flexibility and processing power than you do on your F4 flight controller. On a miniquad it's fine having a slowish processor and one or two UARTs but that doesn't really work on a Rover/Sub. The huge number of PWM outputs, UARTs, and I2C is a must when you're getting into more specialized tasks (like the subs this board focuses on), and that extra flexibility takes space.
In other words, your question is like asking why someone would use a dumptruck when a pickup and a shovel could do the same thing. Different tools for different jobs.
@@OldManCassius thanks!
On the video front, it's perhaps worth noting that Blue Robotics works primarily with tethered ROVs, and the video for those vehicles is transmitted over ethernet via the tether (wifi doesn't work underwater).
Other vehicle types are likely to either not have a live video feed (they may just do onboard recording instead, or have no video at all), or may have an external way of transmitting it (in the same way there's no RC receiver/transmitter or GPS receiver built in - that's unnecessary extra cost for some use-cases that can't make use of those features).
Amazing 🤩 ..
Interesting
Hi! What is the ground robot name?
Hi there! It is a Rover Zero from Rover Robotics: roverrobotics.com/
The only issues is Raspberry Pi 4 are no longer available... Most of the approved vendors listed on the Raspberry web site no longer carry them....
You can buy a Raspberry Pi 4 together with the Navigator 🙂
I can supply you, drop me a line. be quick.
Why is the price so high?
Hi Habiks, the price is high because the components are quite expensive right now due to shortages! We have to pay A LOT more for some of them and keeping the price high helps us keep it available. Hopefully we can make it more affordable eventually!
what about lidar connection?
The Navigator will work with any sensor that is supported in ArduPilot, so there are a number of lidar sensors that can be used!
I'm going to use this to make a drug running drone submarine
youtube algorithm sent me here.... i am not dissapointed..
Very cool, but wow it is expensive.
Chip shortage is a right pain for everyone at the moment, and if it was being sold at unsustainably low prices (compared to current component costs) it would likely end up selling out rapidly and becoming unpurchaseable / having extensive wait times like many other electronics at the moment (including alternative flight controllers).
Hopefully chip shortage calms down soon so prices can reduce to friendlier levels, although unfortunately that likely won't be as soon as any of us would ideally like.
This would have cost thousands a few years ago.
Guys, with this pricing policy you should teach Ilon how to make money. You really want $320 for a board with a couple of sensors? I wish people in my country would buy that for at least a third of the price...
Hi Mike, we totally understand your concern here and we didn't want to sell it for this price either, but with the current electronics supply chain situation it's our only option. We're paying 10x the normal price for some of the components on here and we'd rather have it available, even at a high price, then not available at all! Hopefully that changes in the future!
I just need a cub scout to explain it to me, sry.
Your website is either hacked or is suffering from too aggressive marketing for VPN software. Too bad as I find the product interesting.
Hi Niels, are you saying there are VPN advertisements appearing on our website?
We're aware there are some issues with technical details loading incorrectly at the moment, which we're working to fix, but there definitely shouldn't be any advertisements. If there are we would greatly appreciate any information you can provide about what you saw - please feel free to send us an email at support@bluerobotics.com
Your website is not working
Hi Ognen, we're aware there are some issues right now with loading content correctly, which we're working hard to fix. If there's something else going on we would appreciate if you can send us an email about it at support@bluerobotics.com
Muy buen9
No way I would trust RPi for flight control. I'll stick with microcontrollers.
Ya absolutely not. It's clock speed (rpi4) is slower than the old school f3 controllers.
But let’s say you need to do onboard compute. What would you do? I guess you could slave a flight controller
In the recent years, your prices put your products out of reach for most hobbiests.
Marine technology has traditionally been very expensive. Blue Robotics low prices have changed the industry. Not long ago this would have cost thousands.
@@GeoffCook1000 can't totally agree not totally disagree!. It is a very specialized area of course but, same company used to focus more on the hobbiest. I have a number of the older products back when they took open source seriously. Just going to have to look elsewhere it seems.
Do you know all hobbyists? Have you polled them all to see if most of them think this is too expensive? I don't think it's too expensive for what it does.
@@ZeeCaptainRon well I hope you buy one then. Give a try pricing out the BOM for the pi hat and then you might understand the markup here. Again I do value their contribution but it is overpriced.
The price is too high. Good luck to sell it en masse.
Hope that Pi 4 is more stable that 2-3xx. Older versions were too unpredictable to put them on something flying...
Cost more than navio2, much more
Shame you haven't been able to buy a Raspberry Pi for love nor money anywhere in the world for months now.
You can buy a Raspberry Pi 4 together with the Navigator 🙂
I am afraid i cannot afford it at 320$.
$320 for just a pcb with a bunch of connectors and gpio chip? are you serious? ) RPI PilotPi is only $100
who asked?
@@JoeSheisty-eh5jj have you bought?
@@SK-bl1lp lol yeah i have i was messing with ya but it is super reliable and a great upgrade from the pixhawk. The raspberry pi and the navigator are more organized and much more efficient. It was worth it for our underwater robot
Shame buying a Pi is nearly impossible.
You can opt to get a Pi 4 together with the Navigator :-)
drop me a line, I have limited stock
Shut up and take my money
You can't be selling many at $320 for a HAT.
Complete overpriced ...
Where do you run real-time code?
We don't have a realtime kernel patch but we are running the autopilot code as the highest priority in the scheduler. There's a lot of power and cores in the Raspberry Pi 4 and this hasn't been an issue, but of course it is a valid concern.
@@BlueRobotics I really prefer having a two system approach. It doesn’t cost that much more to add another ARM MCU, and writing real-time state machines is easier than trying to outguess a scheduler. Bonus, each system can watchdog the other.