I started watching the video expecting to hear "I will be selling the plans", which despite how great this robot is, would be a heartbreaker for me. Hearing you say you intend to open source the vital stuff (code/3d files/instructions) while deciding to monetize the nice optional extras was just the sweetest music. I chose to subscribe immediately and will follow this project, and jump into building as soon as you feel the 3d parts are ready to share. You rock. Seriously.
Very nice hexapod. Making it open source, with courses on top of it will help hobbyists or who get into it to just start building. I keep trying to find time outside software development to do some kind of robotics due to burnout, so know what you mean. Now we just need a few more of these and have them dance in sync for your own personal army of spiders :D
@@AecertRobotics It would be cool to add programmable "patrol points" or "animations" (Edit: PRESETS) where you can record a certain movement to a button or two and tap it to play that specific animation. What I was thinking of specifically is that jump animation where instead of sliding the up and down really fast, you can record that to the motion recorder button and have it as a PRESET!! Also, it would be easier to jump and spin, if thats possible, at the same time without having to use the slider!!
@@AecertRobotics I came upon this video exactly because I really want to build a Hexapod that can dance to a beat and has LED lights for a nice party effect :D
just found this channel and I'd take your course in a heart beat. I can learn so much online through various sources but there is such a benefit to a structured curriculum.
Interesting work. Here are some thoughts/ideas I had while watching: 1) Stabilization Muscles: Real-life joints have smaller stabilizer muscles in addition to the macro driver muscles. You might consider adding something along the lines of tensioned ligaments and/or little stabilizer drivers pulling on elastic, spring, pneumatics/hydraulics, or some equivalent tensioning straps across the joints. These could also help provide feedback data to help the system "understand" how the joint is moving by seeing which stabilizers are being tensioned and which ones are relaxed if they are hooked up to a force meter. 2) Force Redirectors: The knee joint, for example, has the kneecap, with muscles attached by ligament to the top and then separate muscles attached to the bottom. This re-directs force "over" the knee joint in such a way to reduce the torque necessary to extend the lower leg and also do so with more stability and less "jittering". That could also help with the "jitter/stutter" issue you're having with the joints. 3) Foot "Angle of Attack" During Walk Cycle: I noticed when showing the example animation for the walk cycle, the foot seemed to simply be "aimed" straight down and then the entire foreleg would slide backwards while perpendicular to the ground. I don't know if this was just for simplicity of animation of if it was actually how it's made to work in the code, but it might be helpful to give it a slight "angle of attack" when it walks; have the leg "tilted" slightly forward so that the tip of the foot points slightly backwards. In other words, the "curve" for the walk cycle shouldn't be a completely straight line, but it should attempt to "dig" ever so slightly down into the ground and there should be a "secondary" cycle for a point somewhere "up-leg", maybe at the middle or even at the "knee" joint with it's own walk cycle calculated just a split-second ahead of the foot. Maybe experiment with it to see if this yields any improvement. 4) Spider Curling: In real-life, Spiders have constant tension on their legs to make them "curl" up underneath them. In order to walk, they actually use their own blood pressure, forcing their circulatory fluid (not technically blood; in spiders it's called _hemolymph)_ out into their legs _against_ that baked-in curling pressure and then "letting up" or "pushing out" to operate the legs. It might be worthwhile to explore this as an option; make some kind of tension system that's constantly trying to fold the legs "under" and then the only real job of the servos is to "extend", and the variable is either "extend harder" or "extend less". It might even be possible to ditch the servo concept altogether, or at least re-purpose it and, instead, use some kind of pneumatic or hydraulic drive to handle the extension of the legs. BTW, this is why spiders curl up like that when they die; no more circulatory pressure means they can't fight against that forced leg curl anymore. 5) Ant GPT: Obviously not even _remotely_ something to implement anytime soon, but perhaps one could incorporate AI functions to do things like optimize leg movements, especially if it integrated aspects like the stabilizer "muscles" with sensors. An AI could take all the info like the tension of each ligament, angle of each joint, power of each servo, accelerometer, body orientation metrics, pneumatic system pressure, etc. and both coordinate the walking calculations and also continually work to optimize them. Hell, even incorporate _vision_ into the system by putting cameras on it. Give it six tiny cameras for good measure and also dramatic irony. They don't necessarily need to see "well", they just need "information" which can be incorporated and utilized for whatever it's worth.
Really cool ideas, ty for the comment! 3 I can and will try, I know exactly what you mean and I think it would help if I get it right. 4 is something I considered. I can do basically that by adding springs to pull the legs down, which would lighten the load on the servos. The more I think about it the more I think I should do it tbh. 5 believe it or not is something I plan on doing kinda. I want another hexapod project where it's walk cycle is completely designed by machine learning. It will have 4 legs, 2 motors in each leg, and I will train it in a simulation with a nueral net on walking forwards and turning. I think it would be a really cool project.
Currently a masters student in Robotics and I have to say your application of robotic kinematics and inverse kinematics to compute joint angles and actually develop the controls for your robot is so cool dude.
Hey man, I think this is seriously cool. Making mistakes is part of learning, and I've heard that one of the best ways to learn is with the intention to teach others. You've inspired me. Thank you.
This is great! To me, I think I could create something similar, but that means I know how much effort was put into this project. Absolutely incredible!
That is amazing!! Doing the modelling in blender blew my mind! I can feel the headaches of doing mechanical parts in it. But it's a great example of improvising with the available tools/knowledge. Congratulations, this was extremely well done!
Thank you so much!! 🙂 Yeah it blew my mind too, but not in a good way 😂 I'm currently remaking the models in fusion360 and it is so much better. Very true! if i wasnt comfortable in blender either i may have been too overwhelmed to even attempt this.
Hello Aecert, my last comment seems not to be saved by RUclips, so i try again: you mentioned that the model for your Hexapod is the MX-Phoenix from Zenta. I also find it the best. It looks great and moves very smoothly and fluidly. I have also been working on a hexapod for the last few years, but my programming skills are very limited, and I have not been able to teach the robot that the world is NOT absolutely flat. That each leg can also touch the ground above or below the baseline. Or maybe not at all. I had been planning to start over when I stumbled across your video. I was immediately intrigued and still am! In Zenta's videos, you can see his Phoenix climbing up and down a large rock or steps. It has a switch built into each foot as a sensor to adjust to the terrain. In the list of your STL files are things like Tibia Switch Mount, Switch Stopper, and Switch Foot, so I assume you at least plan to include switches as well. Which only makes sense if you also use them as sensors. So my question: are you planning something like a Climb Gait, as Zenta seems to have done? That would be great! In general, I'm extremely excited about your way of implementing such a project! I'm looking forward to seeing how it goes.
I am planning on adding the ability for it to walk on uneven terrain!! I'm pretty sure I'll be able to do it to, but I haven't attempted yet. There are switches in the feet, but I haven't even wired them up yet lol. It would most likely be its own mode where one leg moves at a time yeah
This looks like an incredible project I'm a robotics engineer myself and this is peak, would also loved if you would have talked a little about the dynamics of the robot not just the kinematics but it's awesome, hope that the kit and the course sell well the project deserves it.
Thank you so much! That means a ton coming from an actual robotics engineer 😊 What do you mean by "the dynamics" ? Thank you, I really appreciate that ❤️
Thank you so much man, I really appreciate it!! I definitely have been 😄 And this goes without saying, but your projects are amazing and an inspiration ❤️
Brilliant idea! I like the kit idea Especially if you included all the parts and tutorial videos. My daughter is learning robotics and something like this would make a great gift.
I am in ! As long as the price is right. I am just starting in robotics. Bought a official Arduin starter kit. Thansk a lot. Give some stuff for free a get some money in return.... that is the way to go
Oh my God, please do it! I'm so hyped for the open source, the shop and the course! Your hexapod is amazing! I have a dream project to built a hexapod. I know a little bit of electronics, a little bit of arduino, have a 3d printer but I'm not an engineer. This course would help me out and it would be really great!
I am very excited for the course! 100% will get it the day it comes out. Loved the video and I totally want to build the robot because I am in college right now and I want to go into robotics so I really want to learn more about the course. Keep up the great work the video was awesome.
Cool project, looking forward to the kit and course! A nice feature would be to get it running with ROS2. The benefit of this would be, if you put on some sensors like LIDAR, the hexapod would be able to walk autonomous and could use all the features of ROS2.
Thanks, and im really glad to hear that! I 100% agree. Im gonna be honest, Im avoiding ROS atm because i know almost nothing about it, but I used it in college once and it seemed really complex and I had no idea what was going on, I was basically just following instructions and praying it worked. But yeah I hear ya, it would add so many good things. Honestly learning the basics of ROS would be a solid addition to the course, but do you think it would end up being too complex/convoluted? The whole reason im using arduino is to make it as approachable as possible.
@@AecertRobotics yes ROS would be "the next level" as its quite complex to setup. You would also need a raspberry pi and sensors to run it. But if you have a working walk and steering system, i guess its easy to use it with ROS.
I would absolutely love a tutorial. I've been researching hexapods to build one with my sons. Very cool project. Thank you for sharing your experiences.
I really love your videos. I subscribed instantly. Just a beginner question: what code language did you use for project and can you explain how you integrated those equations to the code and I might be sounding dumb, but (as a only python and block coder) I struggled understanding the weighted turning. I would appreciate if you explained it in a video. I hope the best for you and your channel🎉❤🎉.
I would def do the course if it comes out, I would love to learn how all this leg placement stuff works, as I'm currently building a robot arm which would really benefit from it. 10/10 channel btw.
this video just got recommended to me and man, it's fantastic and the jumping surprised me. lol also, i can totally relate with the burnout from web dev work.
Two ideas I have. One some sort of mount plate on top for attaching accessories. Two wireless charger on the bottom so it can walk onto and set itself down on a charging pad. I think just these two modifications would greatly expand the use cases for this awesome robot. Dang it, now I want to build one.
Ooooh a wireless charging base is a really good idea!! It doesnt even have to be wireless tbh... Ill have to look into this. Accessories are in the pipeline 😉
Hey! this is awesome! I am a robotics engineering student and I must say you did a great job there. I will buy your course the day it comes out! please also include pcb design in the course. Thanks!
Had same jitter 7:20 on the loosest leg of mine. I wouldn't really blame the servo as such. It's simply the mass of the leg at a distance and the speed of the movement causing overshoot on the servo position so it tries to correct back the other way and the same thing happens and basically it resonates. The servo is kinda doing what it is supposed to be doing but I appreciate that doesn't get you any closer to getting rid of the problem. Indeed increasing resistance to movement on the leg is probably the best bet for a fix.
I just got done building a quadruped one with the cheap sg90 servos and it's cool but wow yours is so smooth. I can't get much power out of mine without it shutting down. Definitely going to follow your project, thanks.
Thanks! Oooh great job! Im interested in trying to make a hexapod from mg90s at some point. It sounds either your battery is too weak, or a part of your electronics cant handle the amps required.
@@AecertRobotics I used a Arduino nano and a shield board, I'm guessing since it's being powered from one source is the culprit as I'm using a 7.4v 1300mah lipo and a buck converter to knock down the voltage to 5v, tried upping the voltage and it doesn't like going much higher than about 6v.
I would love to see a Walker like this interact with the world around it. This means if it sees a step or rock in the way it can intuitively lift its legs in a way to get over the obstacle without having the “pilot” imputing a command to do so.
That sort of computer vision AI is above my head unfortunately. What I do plan on doing is utilizing the sensors in the feet to detect when a foot is on the ground and balance/walk accordingly.
Thanks! Yeah thats a big reason why i chose this project, it was a really fun challenge. The key is to break the main problem (making a hexapod walk) up into smaller more manageable problems.
Amazing project! It's nice to see a tutorial about this robot. In my opinion, it's much more interesting to observe the code and mathematics principles of hexapod in video format (not for free, of course). I mean, to make paid video course with explanation of principles of walking, moving, and etc. Good luck😁👍🔥
Thank you!! And I totally agree. I havent put too much thought into it yet, but I do know the course will definitely have a ton of videos! Thanks lmao ill need it 😂
Hey there! Just watched your hexapod video and I'm blown away by your amazing creation! 🤯 Your skills as a programmer and game developer truly shine in this project. I've got a few ideas that could make your hexapod even cooler, and I definitely think you should consider open-sourcing it to get the community involved! Let me flow here for a second.. my brain is literally gushing with ideas: 1. Digital replica in a game engine: With your game dev background, you could create a virtual version of the hexapod in Unity or Unreal. This way, we can all experiment with AI algorithms, locomotion strategies, and obstacle navigation in a virtual environment before implementing them in the real robot; even people without the soldering skills or 3D printers could collaborate on making the control software better. 2. Fisheye lens cameras with low-resolution but super-fast sensors: Adding these cameras on the top and bottom of the pod's body would provide a 360-degree view of its surroundings. The low-res, high-speed sensors would help the robot quickly process its environment and make smarter decisions. (already thinking about implementing a really slim neural net for basic sensory perception and a dedicated vision system for object recognition) 3. Machine learning for advanced movement & navigation: Train the hexapod to adapt to different terrains and obstacles using ML algorithms. It could learn to climb, walk on uneven surfaces, and maybe even perform some sweet acrobatic moves! 4. Crowdsource AI development: Open-sourcing your project means the entire community could contribute their own AI algorithms or improvements for the hexapod, pushing its intelligence to new heights! 5. Gaming & entertainment: Create a game where players control their own hexapods in a virtual arena, battling other robots or completing challenges, controlled by AI.. maybe even with different checkpoint files (of limited size, obviously.. not everyone has a A100 laying around 🤣) of trained networks being able to be uploaded to a website where they would be set up against other neural networks in battle. A hexapod would be perfect for this! .... ok, ok, maybe I'm rambling off a bit on this one, but still 😁 Your hexapod project is seriously impressive, and I believe open-sourcing it would inspire others to contribute and learn from your work. The possibilities are endless! Keep up the fantastic work, and I'm excited to see where this project goes! 🚀🕷🎮
Thank you so much 😊 I love all these ideas! 1. ive thought a lot about this and I really want to do it. My only issue with this is I would want to make it so I can just take the code from the game engine and run it on the hexapod, but to be able to do that will take a ton of work. But even just getting it working in an engine would be sick and I do want to do it. 2. I do want to add a camera. Fisheye lense purely for AI is an interesting idea. Do you think the legs getting into view would mess things up? 3. This would be amazing. I did have the idea of trying to train a neural network to make the hexapod walk, but for this I would definitelyyy need the simulation working. Training it to navigate and walk on uneven surfaces should be possible, but is way over my head. 4. The code will definitely be open sourced! i look forward to what the community comes up with. 5. Im not sure about that.... XD But i was thinking more along the lines of a 1v1 hexapod battle where each hexapod has a laser gun attachment, and you try to hit the other hexapod to lower its health, and at 0 health it shuts down. Im not sure how laser tag tech works, but i dont think it would be that hard to do and would be really fun! Thank you so much man, I really appreciate it ❤️
This is really cool! You also might be able to get rid of the jittering legs by adding a pid controller for the motors, but honestly the friction plates are a lot less work and perform fine.
Thank you! Yeah i tried pretty much everything except a dedicated servo controller board, but im trying really hard to avoid adding complexity to the electronics unless necessary. But honestly I should try a servo controller combined with a more powerful arduino mega replacement, because it might be a lot less complicated than I think.
@@AecertRobotics If you have enough I/O pins, servo controllers are just unneeded complexity. I don't understand why so many people use them. The jitter is a mechanical issue. A "pid controller" isn't going to fix them. My biggest complaint about most servos is they use potentiometers. Potentiometers in robot projects are a bad idea. Potentiometers wear out over time. I'd love to find a cheap way to add hall sensors to servos.
Very cute hexapod. Since this is your first robotics project in a long while and you changed servos twice, maybe hold off on course and kit until you have enough experience to show the optimal way to do things and include the best parts for the project
I think you're the first person to call it cute 😄 100% agree. I would not feel comfortable releasing anything until I can make sure the servos won't be blowing with normal use.
Awesome stuff I just thought 🤔 about hex pod ii No remote control however sensors to move out of the , respond to sound, and or instructions etc You have obviously done the hardest part of it and it looks like a winner 🏆
Sweet hexapod! Servos can be a pain to deal with. I've seen a lot of people praise the Dynamixel servos but they're pricey. It could be worth looking at though. This would be a great base to add an airsoft gun to and make it a Mech Warfare type bot. Or go optical and use something like FPV Combat. Lots of possibilities! I'm very tempted to build one once you've posted files and code. Can't wait to see more of this!
Thanks! Dynamixel are better in every way except price, but price trumps all... Oh lord airsoft would be actually scary. I was thinking like a high powered nerf gun. I loveeee the idea of slapping a camera on it and driving it remotely. Really happy to hear that!!
@@AecertRobotics Nerf would be great too! Better than airsoft BBs all over the place. There's some 3D printed electric airsoft mechanisms that use brushless motors for some serious power. Analog FPV is relatively cheap. Goggles are the most expensive part. Could do a cheap receiver to a TV though. May want to look at a raspberry pi zero w, mjpeg streamer to stream video over wifi and php to control it remotely. 👍
subscribed. Would be thrilled if there will be a detailed tutorial and files for this project so that we can try it too, but of course no pressure on you because I understand how time and effort consuming that would be. anyway, good job!
I started watching the video expecting to hear "I will be selling the plans", which despite how great this robot is, would be a heartbreaker for me. Hearing you say you intend to open source the vital stuff (code/3d files/instructions) while deciding to monetize the nice optional extras was just the sweetest music. I chose to subscribe immediately and will follow this project, and jump into building as soon as you feel the 3d parts are ready to share. You rock. Seriously.
Makes me really happy to hear this 🙂 Thank you 😊
(Also if you want to take an early look at the files email me and ill send them to you!)
@@AecertRoboticscan you plz share you email id
@@parthtailor1189 check my RUclips homepage under about!
I agree. These kind of acts is what make human kind evolve.
@@AecertRoboticsHi, I am a student and this project seems interesting. Can I also have an early look at the files?
Obviously, we all want a tutorial! Srsly.
is that not what this is?
I demand a Tutorial!!!
Yes def
TUTORIAL!!!
I definitely would like one! I've always wanted to build one of these!
Very nice hexapod. Making it open source, with courses on top of it will help hobbyists or who get into it to just start building. I keep trying to find time outside software development to do some kind of robotics due to burnout, so know what you mean. Now we just need a few more of these and have them dance in sync for your own personal army of spiders :D
Thank you! And yeah I totally agree.
Omg that would be hilarious :D I can confirm I do plan on adding dance moves :)
@@AecertRobotics It would be cool to add programmable "patrol points" or "animations" (Edit: PRESETS) where you can record a certain movement to a button or two and tap it to play that specific animation. What I was thinking of specifically is that jump animation where instead of sliding the up and down really fast, you can record that to the motion recorder button and have it as a PRESET!!
Also, it would be easier to jump and spin, if thats possible, at the same time without having to use the slider!!
@@AecertRobotics I came upon this video exactly because I really want to build a Hexapod that can dance to a beat and has LED lights for a nice party effect :D
The fact that you're willing to release everything for free (after all this insane effort and time) makes me respect you so much.
I would purchase the course in a heartbeat! You’ve earned my subscription and look forward to further updates!
just found this channel and I'd take your course in a heart beat. I can learn so much online through various sources but there is such a benefit to a structured curriculum.
You are a class act man, open-sourcing this, really helpful for those of us who can tinker but have never built a hexapod. Liked and subbed :)
Interesting work. Here are some thoughts/ideas I had while watching:
1) Stabilization Muscles: Real-life joints have smaller stabilizer muscles in addition to the macro driver muscles. You might consider adding something along the lines of tensioned ligaments and/or little stabilizer drivers pulling on elastic, spring, pneumatics/hydraulics, or some equivalent tensioning straps across the joints. These could also help provide feedback data to help the system "understand" how the joint is moving by seeing which stabilizers are being tensioned and which ones are relaxed if they are hooked up to a force meter.
2) Force Redirectors: The knee joint, for example, has the kneecap, with muscles attached by ligament to the top and then separate muscles attached to the bottom. This re-directs force "over" the knee joint in such a way to reduce the torque necessary to extend the lower leg and also do so with more stability and less "jittering". That could also help with the "jitter/stutter" issue you're having with the joints.
3) Foot "Angle of Attack" During Walk Cycle: I noticed when showing the example animation for the walk cycle, the foot seemed to simply be "aimed" straight down and then the entire foreleg would slide backwards while perpendicular to the ground. I don't know if this was just for simplicity of animation of if it was actually how it's made to work in the code, but it might be helpful to give it a slight "angle of attack" when it walks; have the leg "tilted" slightly forward so that the tip of the foot points slightly backwards. In other words, the "curve" for the walk cycle shouldn't be a completely straight line, but it should attempt to "dig" ever so slightly down into the ground and there should be a "secondary" cycle for a point somewhere "up-leg", maybe at the middle or even at the "knee" joint with it's own walk cycle calculated just a split-second ahead of the foot. Maybe experiment with it to see if this yields any improvement.
4) Spider Curling: In real-life, Spiders have constant tension on their legs to make them "curl" up underneath them. In order to walk, they actually use their own blood pressure, forcing their circulatory fluid (not technically blood; in spiders it's called _hemolymph)_ out into their legs _against_ that baked-in curling pressure and then "letting up" or "pushing out" to operate the legs. It might be worthwhile to explore this as an option; make some kind of tension system that's constantly trying to fold the legs "under" and then the only real job of the servos is to "extend", and the variable is either "extend harder" or "extend less". It might even be possible to ditch the servo concept altogether, or at least re-purpose it and, instead, use some kind of pneumatic or hydraulic drive to handle the extension of the legs. BTW, this is why spiders curl up like that when they die; no more circulatory pressure means they can't fight against that forced leg curl anymore.
5) Ant GPT: Obviously not even _remotely_ something to implement anytime soon, but perhaps one could incorporate AI functions to do things like optimize leg movements, especially if it integrated aspects like the stabilizer "muscles" with sensors. An AI could take all the info like the tension of each ligament, angle of each joint, power of each servo, accelerometer, body orientation metrics, pneumatic system pressure, etc. and both coordinate the walking calculations and also continually work to optimize them. Hell, even incorporate _vision_ into the system by putting cameras on it. Give it six tiny cameras for good measure and also dramatic irony. They don't necessarily need to see "well", they just need "information" which can be incorporated and utilized for whatever it's worth.
Really cool ideas, ty for the comment!
3 I can and will try, I know exactly what you mean and I think it would help if I get it right.
4 is something I considered. I can do basically that by adding springs to pull the legs down, which would lighten the load on the servos. The more I think about it the more I think I should do it tbh.
5 believe it or not is something I plan on doing kinda. I want another hexapod project where it's walk cycle is completely designed by machine learning. It will have 4 legs, 2 motors in each leg, and I will train it in a simulation with a nueral net on walking forwards and turning. I think it would be a really cool project.
Saw this on Reddit and I'm happy that RUclips served this up. I appreciate all your documentation and instruction. Instant Sub from me!
Super glad you found my channel! Really appreciate the sub, there will be much more to come :)
Kit+course idea sound really good
Amazing project, the movement of the robot is so smooth. Congrats! If you can share the porjects as open source, it will be amazing!
Thank you so much! I 100% plan on it!
Currently a masters student in Robotics and I have to say your application of robotic kinematics and inverse kinematics to compute joint angles and actually develop the controls for your robot is so cool dude.
That means a lot man I really appreciate it 🙂
Can't wait for the course on this project
to see such artifact creature, brings me such joy that it kinda makes the sadness go away. Thank you!
You got my sub from the sentence that ended 'in blender' :) Finally someone else that uses blender for precision modeling!
Hey man, I think this is seriously cool. Making mistakes is part of learning, and I've heard that one of the best ways to learn is with the intention to teach others. You've inspired me. Thank you.
Thanks dude :)
I completely agree! You gain a much better understanding of something when you have to teach it to someone else.
I recognize this! I saw your update videos on reddit, and you even answered one of my questions that I asked you! Nice project
Thank you!! Im glad you found your way here 🙂
You have done this really well
Thank you so much
This is great! To me, I think I could create something similar, but that means I know how much effort was put into this project. Absolutely incredible!
That is amazing!! Doing the modelling in blender blew my mind! I can feel the headaches of doing mechanical parts in it. But it's a great example of improvising with the available tools/knowledge.
Congratulations, this was extremely well done!
Thank you so much!! 🙂 Yeah it blew my mind too, but not in a good way 😂 I'm currently remaking the models in fusion360 and it is so much better.
Very true! if i wasnt comfortable in blender either i may have been too overwhelmed to even attempt this.
Hello Aecert,
my last comment seems not to be saved by RUclips, so i try again:
you mentioned that the model for your Hexapod is the MX-Phoenix from Zenta. I also find it the best. It looks great and moves very smoothly and fluidly. I have also been working on a hexapod for the last few years, but my programming skills are very limited, and I have not been able to teach the robot that the world is NOT absolutely flat. That each leg can also touch the ground above or below the baseline. Or maybe not at all. I had been planning to start over when I stumbled across your video. I was immediately intrigued and still am!
In Zenta's videos, you can see his Phoenix climbing up and down a large rock or steps. It has a switch built into each foot as a sensor to adjust to the terrain.
In the list of your STL files are things like Tibia Switch Mount, Switch Stopper, and Switch Foot, so I assume you at least plan to include switches as well. Which only makes sense if you also use them as sensors. So my question: are you planning something like a Climb Gait, as Zenta seems to have done? That would be great!
In general, I'm extremely excited about your way of implementing such a project! I'm looking forward to seeing how it goes.
I am planning on adding the ability for it to walk on uneven terrain!! I'm pretty sure I'll be able to do it to, but I haven't attempted yet. There are switches in the feet, but I haven't even wired them up yet lol.
It would most likely be its own mode where one leg moves at a time yeah
This looks like an incredible project I'm a robotics engineer myself and this is peak, would also loved if you would have talked a little about the dynamics of the robot not just the kinematics but it's awesome, hope that the kit and the course sell well the project deserves it.
Thank you so much! That means a ton coming from an actual robotics engineer 😊
What do you mean by "the dynamics" ?
Thank you, I really appreciate that ❤️
Looking nice 👍 and good luck with your future project making it a kit with tutorials. Have fun!
Thank you so much man, I really appreciate it!! I definitely have been 😄
And this goes without saying, but your projects are amazing and an inspiration ❤️
I really like the idea of the course!
Beautiful project! All the best!
Love the idea of the course.
This is awesome and needs more likes. I want a whole stock of these for when I want my mail brought to my room and for when I fight Spider-Man.
this is amazing and so beautifully done , cant wait to see the actual tutorial
Brilliant idea! I like the kit idea Especially if you included all the parts and tutorial videos. My daughter is learning robotics and something like this would make a great gift.
Thank you so much!! That is exactly the plan, and I completely agree!
I am in ! As long as the price is right. I am just starting in robotics. Bought a official Arduin starter kit. Thansk a lot. Give some stuff for free a get some money in return.... that is the way to go
Oh my God, please do it! I'm so hyped for the open source, the shop and the course! Your hexapod is amazing! I have a dream project to built a hexapod. I know a little bit of electronics, a little bit of arduino, have a 3d printer but I'm not an engineer. This course would help me out and it would be really great!
You are exactly the type of person the kit plus course is for 🙂 I can't wait for it's release!
I like the idea!, looking forward :D
honestly cant stop watching this video the smoothness of this thing is insane
I'd definitely be interested in a kit and a course! I've always wanted to learn more about these little robot projects and how to do them as a hobby.
wow this is so cool! and the aesthetic of the 3d design too
Thank you so much! That was a very important consideration of mine when designing it.
The moment i heard this is open source I subscribed. Need more people like you!
I am very excited for the course! 100% will get it the day it comes out. Loved the video and I totally want to build the robot because I am in college right now and I want to go into robotics so I really want to learn more about the course. Keep up the great work the video was awesome.
I'm so happy to hear that!! Thank you so much 🙂
Keep this up and this channel will get bigger
Amazing video! Deserves a lot more views for all the effort
Thank you so much! :)
well done , your progress and approach to this looks gr8 , subscribed , will defo be following this project
OMG, it looks so smooth
Cool project, looking forward to the kit and course! A nice feature would be to get it running with ROS2. The benefit of this would be, if you put on some sensors like LIDAR, the hexapod would be able to walk autonomous and could use all the features of ROS2.
Thanks, and im really glad to hear that!
I 100% agree. Im gonna be honest, Im avoiding ROS atm because i know almost nothing about it, but I used it in college once and it seemed really complex and I had no idea what was going on, I was basically just following instructions and praying it worked.
But yeah I hear ya, it would add so many good things. Honestly learning the basics of ROS would be a solid addition to the course, but do you think it would end up being too complex/convoluted? The whole reason im using arduino is to make it as approachable as possible.
@@AecertRobotics yes ROS would be "the next level" as its quite complex to setup. You would also need a raspberry pi and sensors to run it. But if you have a working walk and steering system, i guess its easy to use it with ROS.
Would 100% buy the course as you describe it.
Great project and implimentation. I'm looking to make a spider or hexpod robot soon. So it's great to get these hints on how you did it.
Thank you so much! That's awesome! If you have any specific questions don't hesitate to ask
This is great! I'd love to see a tutorial! Keep up the innovative content!
I would absolutely love a tutorial. I've been researching hexapods to build one with my sons.
Very cool project. Thank you for sharing your experiences.
Awww thats great 🙂 I can't wait to release the course!
I need a tutorial so much!
Nice job!
I really love your videos. I subscribed instantly. Just a beginner question: what code language did you use for project and can you explain how you integrated those equations to the code and I might be sounding dumb, but (as a only python and block coder) I struggled understanding the weighted turning. I would appreciate if you explained it in a video. I hope the best for you and your channel🎉❤🎉.
I would def do the course if it comes out, I would love to learn how all this leg placement stuff works, as I'm currently building a robot arm which would really benefit from it. 10/10 channel btw.
Yes! I want to print and build this. A tutorial and sharing code would be so amazing. Great job !
I'm incredibly enthusiastic to see this project and to start robotics for myself after a long time procrastinating.
Hell yeah dude! Best time to start is right now 🙂
I love this! And yes I would buy both the kit and the lesson!
this video just got recommended to me and man, it's fantastic and the jumping surprised me. lol
also, i can totally relate with the burnout from web dev work.
Once I start my job in cybersecurity I'd gladly buy your kit and your course!
it is really interesting, and im excited and waiting for tutorials 👉👈
Two ideas I have. One some sort of mount plate on top for attaching accessories. Two wireless charger on the bottom so it can walk onto and set itself down on a charging pad. I think just these two modifications would greatly expand the use cases for this awesome robot. Dang it, now I want to build one.
Ooooh a wireless charging base is a really good idea!! It doesnt even have to be wireless tbh... Ill have to look into this.
Accessories are in the pipeline 😉
Hey! this is awesome! I am a robotics engineering student and I must say you did a great job there. I will buy your course the day it comes out! please also include pcb design in the course. Thanks!
Thank you so much!! The course will cover everything 😉 (although I do plan on releasing a basic pcb tutorial on RUclips)
It´s look very awesome dude! Clap
Awesome stuff bro
Had same jitter 7:20 on the loosest leg of mine. I wouldn't really blame the servo as such. It's simply the mass of the leg at a distance and the speed of the movement causing overshoot on the servo position so it tries to correct back the other way and the same thing happens and basically it resonates. The servo is kinda doing what it is supposed to be doing but I appreciate that doesn't get you any closer to getting rid of the problem. Indeed increasing resistance to movement on the leg is probably the best bet for a fix.
100% agree. For how inexpensive the servo is I'm pretty happy with it tbh
"It jumps!"...lol, thumbs up earned. Great build
Thanks! 😄
Sehr toll gemacht!!👍
Yes, tutorial and kit 👍
I subscribd! This is crazy and epic! I am about to watch your next video
Amazing work!
Yes, definitely want a tutorial. Need one better than the ant one.
New robotics channel 🎉
this is the most satisfying but terrfying also inspiring thing ive ever seen
Can’t wait for that hopefully sooner
Hey Aecert, this is top tier content 👌. I've been looking to simulate something like this before actually purchasing the parts. Looks awesome.
Thank you so much :) Dude i regret not building this in a simulation beforehand tbh. Would you use ROS or do it from scratch?
I just got done building a quadruped one with the cheap sg90 servos and it's cool but wow yours is so smooth. I can't get much power out of mine without it shutting down. Definitely going to follow your project, thanks.
Thanks! Oooh great job! Im interested in trying to make a hexapod from mg90s at some point. It sounds either your battery is too weak, or a part of your electronics cant handle the amps required.
@@AecertRobotics I used a Arduino nano and a shield board, I'm guessing since it's being powered from one source is the culprit as I'm using a 7.4v 1300mah lipo and a buck converter to knock down the voltage to 5v, tried upping the voltage and it doesn't like going much higher than about 6v.
Excellent video, verry educational and beneficial to everyone.
I would love to see a Walker like this interact with the world around it. This means if it sees a step or rock in the way it can intuitively lift its legs in a way to get over the obstacle without having the “pilot” imputing a command to do so.
That sort of computer vision AI is above my head unfortunately. What I do plan on doing is utilizing the sensors in the feet to detect when a foot is on the ground and balance/walk accordingly.
love this project
Great work !
Great build 👏,, Figuring out joint kinematics in this complex build in itself is the USP of this project and the upcoming course.
Thanks! Yeah thats a big reason why i chose this project, it was a really fun challenge. The key is to break the main problem (making a hexapod walk) up into smaller more manageable problems.
There should be a hexapod war competitino live streamed on ESPN.
8:49 jumping hexapod is awesome
very cool, would be awesome to see other people try to code it and get it to leap forward
also the strafe mode and drive mode could potentially be different joysticks
Dang dude this is a good idea... I was thinking a jump button, but i didnt consider making the jump leap in the direction of your joystick.
This is really damn cool. Well done!
Thank you so much!
This is really great. I do lots of 3dprinting but am now trying to learn electronics so I'd be very interested in the course.
Thanks! I feel like this would be a great project to learn on so im glad to hear that
bro, deserved a sub
Amazing project!
It's nice to see a tutorial about this robot. In my opinion, it's much more interesting to observe the code and mathematics principles of hexapod in video format (not for free, of course). I mean, to make paid video course with explanation of principles of walking, moving, and etc.
Good luck😁👍🔥
Thank you!! And I totally agree. I havent put too much thought into it yet, but I do know the course will definitely have a ton of videos! Thanks lmao ill need it 😂
Its happy jumps made me happy!
What interesting motor skills! 🤔 like a real spider moved ! 👍
If you release this ill happily make one. Id probably replace the arduino with a stm32 or esp32.
I can't wait to see how people alter this and expand upon it
Hey there! Just watched your hexapod video and I'm blown away by your amazing creation! 🤯 Your skills as a programmer and game developer truly shine in this project. I've got a few ideas that could make your hexapod even cooler, and I definitely think you should consider open-sourcing it to get the community involved! Let me flow here for a second.. my brain is literally gushing with ideas:
1. Digital replica in a game engine: With your game dev background, you could create a virtual version of the hexapod in Unity or Unreal. This way, we can all experiment with AI algorithms, locomotion strategies, and obstacle navigation in a virtual environment before implementing them in the real robot; even people without the soldering skills or 3D printers could collaborate on making the control software better.
2. Fisheye lens cameras with low-resolution but super-fast sensors: Adding these cameras on the top and bottom of the pod's body would provide a 360-degree view of its surroundings. The low-res, high-speed sensors would help the robot quickly process its environment and make smarter decisions. (already thinking about implementing a really slim neural net for basic sensory perception and a dedicated vision system for object recognition)
3. Machine learning for advanced movement & navigation: Train the hexapod to adapt to different terrains and obstacles using ML algorithms. It could learn to climb, walk on uneven surfaces, and maybe even perform some sweet acrobatic moves!
4. Crowdsource AI development: Open-sourcing your project means the entire community could contribute their own AI algorithms or improvements for the hexapod, pushing its intelligence to new heights!
5. Gaming & entertainment: Create a game where players control their own hexapods in a virtual arena, battling other robots or completing challenges, controlled by AI.. maybe even with different checkpoint files (of limited size, obviously.. not everyone has a A100 laying around 🤣) of trained networks being able to be uploaded to a website where they would be set up against other neural networks in battle. A hexapod would be perfect for this! .... ok, ok, maybe I'm rambling off a bit on this one, but still 😁
Your hexapod project is seriously impressive, and I believe open-sourcing it would inspire others to contribute and learn from your work. The possibilities are endless! Keep up the fantastic work, and I'm excited to see where this project goes! 🚀🕷🎮
Thank you so much 😊 I love all these ideas!
1. ive thought a lot about this and I really want to do it. My only issue with this is I would want to make it so I can just take the code from the game engine and run it on the hexapod, but to be able to do that will take a ton of work. But even just getting it working in an engine would be sick and I do want to do it.
2. I do want to add a camera. Fisheye lense purely for AI is an interesting idea. Do you think the legs getting into view would mess things up?
3. This would be amazing. I did have the idea of trying to train a neural network to make the hexapod walk, but for this I would definitelyyy need the simulation working. Training it to navigate and walk on uneven surfaces should be possible, but is way over my head.
4. The code will definitely be open sourced! i look forward to what the community comes up with.
5. Im not sure about that.... XD But i was thinking more along the lines of a 1v1 hexapod battle where each hexapod has a laser gun attachment, and you try to hit the other hexapod to lower its health, and at 0 health it shuts down. Im not sure how laser tag tech works, but i dont think it would be that hard to do and would be really fun!
Thank you so much man, I really appreciate it ❤️
i would pay for that course. I'm subscribed and i hope to see that materialize soon
Would definitely buy your course
YES! I'd love to see a tutorial; I would like to learn as well and possibly improve it!
Excellent!
This is really cool! You also might be able to get rid of the jittering legs by adding a pid controller for the motors, but honestly the friction plates are a lot less work and perform fine.
Thank you! Yeah i tried pretty much everything except a dedicated servo controller board, but im trying really hard to avoid adding complexity to the electronics unless necessary. But honestly I should try a servo controller combined with a more powerful arduino mega replacement, because it might be a lot less complicated than I think.
@@AecertRobotics If you have enough I/O pins, servo controllers are just unneeded complexity. I don't understand why so many people use them.
The jitter is a mechanical issue. A "pid controller" isn't going to fix them.
My biggest complaint about most servos is they use potentiometers. Potentiometers in robot projects are a bad idea. Potentiometers wear out over time. I'd love to find a cheap way to add hall sensors to servos.
I like this channel already
Very cute hexapod.
Since this is your first robotics project in a long while and you changed servos twice, maybe hold off on course and kit until you have enough experience to show the optimal way to do things and include the best parts for the project
I think you're the first person to call it cute 😄
100% agree. I would not feel comfortable releasing anything until I can make sure the servos won't be blowing with normal use.
Awesome stuff
I just thought 🤔 about hex pod ii
No remote control however sensors to move out of the , respond to sound, and or instructions etc
You have obviously done the hardest part of it and it looks like a winner 🏆
Fully ai + voice control sounds really freaking cool tbh
Nice jump!!
You better scare some kids on Halloween with this thing. Great work!
That would be hilarious 😄
Sweet hexapod! Servos can be a pain to deal with. I've seen a lot of people praise the Dynamixel servos but they're pricey. It could be worth looking at though. This would be a great base to add an airsoft gun to and make it a Mech Warfare type bot. Or go optical and use something like FPV Combat. Lots of possibilities! I'm very tempted to build one once you've posted files and code. Can't wait to see more of this!
Thanks!
Dynamixel are better in every way except price, but price trumps all...
Oh lord airsoft would be actually scary. I was thinking like a high powered nerf gun. I loveeee the idea of slapping a camera on it and driving it remotely.
Really happy to hear that!!
@@AecertRobotics Nerf would be great too! Better than airsoft BBs all over the place. There's some 3D printed electric airsoft mechanisms that use brushless motors for some serious power. Analog FPV is relatively cheap. Goggles are the most expensive part. Could do a cheap receiver to a TV though. May want to look at a raspberry pi zero w, mjpeg streamer to stream video over wifi and php to control it remotely. 👍
subscribed. Would be thrilled if there will be a detailed tutorial and files for this project so that we can try it too, but of course no pressure on you because I understand how time and effort consuming that would be. anyway, good job!
That's the plan! And ty 🙂
I need the course