It does a good job, however needs more damping indeed. Switching point of interest instead of accurately following it is what's needed. And personally I'm used to a signature steady wide open shot with a lot of machinery in the frame. Change is not always good. A lot of videos today are dynamic and shot from a handheld cam. James style stands apart and that is good IMO.
@@kittguy And another bunch of videos are filmed on a tripod, with wide shots. James' style wasn't special at all, in fact a robot filming him IS a special style he needs. Just needs some tuning, that's all.
It's a VERY good machine to carry a quite massive load and for keeping it level and stable. But for it carry you a beer while you're feeling lazy... well I think it would be warm by the time it got there lol. You did mention that it was overly complicated, and I agree... for SIMPLE tasks. But it really is the perfect design for what it was based on (cleaning the stairs)
If I were to do this project, I would cheat and mount a track on the wall that the robot could attach to and drive itself up the wall like one of those stair-lift-chair things. Or attach and detach itself to an installed winch or crane. That would only help for a home use that's planned ahead of time like that - not really as robust as a solution as this robot that could go up many different types of stairs. EDIT 12:32 in the video he has the same idea haha I wrote this comment before finishing the vid
The new camera rig seems like it's working out great, but with this project, because you've already added the 2kg weight shifting element, why not just use two sets of balancing wheels and a single actuator between them?
Funny - I had the opposite thought - since it's using 6 wheels, why have a shifting mass when you could just add a static mass to the middle section? :P
I'm pretty sure sensors were limit switches when I was learning robotics back in the 80s. There were infra-red sensors too, but there's nothing wrong with an, *ahem* _simple touch sensor_ AKA limit switch. :)
Love the multiple different types of pullies and different engineering designs in this. On a quick note you should try to use mecanum wheels, they give lots of control when used right and you can go in almost any direction you want!
What would be super cool is if someone could make like a rail system around a house that bots can attach to and ride around. For the stairs the rail could go up a wall and then across the ceiling out the way.
I gotta say I switched to Onshape about a year and a half ago when I switched to Linux and coming from fusion I gotta say I love it and I love to see them sponsoring you
nice automatism . You can also try triangular wheels (one wheel on each vertices), all three powered by a central sprocket , and the triangle turns and grips . Like a powered climbing stairs cart
🤣 You think a university would only take "years"? 😂😂 An academic would sit on a problem like for his/her/its entire career. And in that time he/she/it would build a medium size research group, obtain squillions in taxpayer funded grants to fund said group, reach full professor relatively early compared to a more esoteric academic colleague and then spend the rest of his/her/its career dining out on his/her/its one pony trick without any of his/her/its sponsors ever really achieving a commercially viable product. Even more sadly he/she/it will successfully supervise higher degrees and they too will follow in his/her/its path at his/her/its institue or some other. And so and so on and so on. 🤣😂🤣
Fantastic as always. I thought that seven minutes to get up the stairs was quite impressive. :) A few weeks back I thought that your motion camera rig was another amazing project. However, when you use it now, there are parts of the video (like the closing section here) where I have to look away as they make me fill quite ill very quickly -- the mimicking of a drifting hand-held camera with no stabilization is too good! :) But this may be just me.
If the compartment of the battery can move to accommodate for the change of weight in the direction needed, that's a huge win for efficiency. Now imagine being able to do this with more than one battery separately for even more control and capacity.
I know for a FACT you used your camera bot and damnnnnnnn that lil homie was build by a pro cause it didn’t miss a shot. Love the video have a great day!!
I see you use your camera robot for filming now. Thats pretty awesome that you get something really useful out of your inventions as well from time to time.
Awesome job man. One suggestion. Instead of shifting weight, you can use flywheels. You can control their speed and direction to adjust the angular momentum. I think it will be comparatively simple and compact.
Great video! I had a thought though, and I'm not sure if I'm missing something (in which case, please correct me), but if you have the moving mass that can change where it balances, why couldn't it be just two segments? Like, it puts one segment up one step, shifts weight onto that one, lifts the back segment up, transfers weight onto that one, etc.
I think it only needs to drive the front pair of wheels when the middle pair is at the same level as those front pair. This means you could simplify by having a simple gear or friction drive that takes the drive off the middle pair and supplies it to the front pair when they are at the same level?
a couple rails with a section of teeth in the middle would make it easier for a small robot to quickly climb the stairs. depending on size you could get it to corner on the rail and even a weak, slow climb would still probably be faster than all the stop and go with this bot. this robot is amazing in its own way and if it was only a step or two it would be fine but 7 min is a while and then you need a good bit of sensors to keep it from just driving off the edge.
At least, I think you can automate the stair climb with 1 sonar sensor on the front: You still have control of the driving of course. You can drive it up to a step, and hit a button that will start the stair climbing (maybe with the same button to stop the auto-climb so you can take over too). The program will first make sure it's close enough to the step, maybe even bumping into the step. It will then raise the front section until the sensor doesn't detect a close object. It then drives forward a set distance, stopping if the sensor gets too close to the next step (if there is one). I'm not going to write out the whole thing, but I think you can get the idea for the algo from this. Just one sensor needed! :)
If you can find it have a look at the British Army's "wheelbarrow" from the 1980s. It is an EOD robot that carries tools and can negotiate stairs no problem. It uses tracks and has a hull that slides fore and aft for centre of gravity control. Well done on this design...
Not bad for a first step... I'll see myself out now. In all seriousness though, it reminds me of the way a baby climbs stairs. Functional, effective, but very slow. Very interested to see what else comes out.
How about a hopping / pogo stick robot that can hop up the steps? Would need some kind of active damping on the beer carrying arm though. Sounds right up your alley with some reaction wheals and or gyro's for stability and inducing forward motion and rotation etc
Wouldn't you simply drink the beer before climbing the stairs? And what is so important at the top of the stairs anyway? Tell her to come down, you're too busy watching the cricket
Always enjoy seeing your solution to hese problems. Adding a payload area seems like it might get dicey. You're already fighting balance issues. Seems like figuring out additional weight distribution could be a challenge. Thanks for sharing!
Rather than sensors for detecting the distance to the next step couldn't you use switches on each side (front left and front right) and let the robot bump into the stair like the early roombas did to detect the stair and then orient itself correctly (straighten out and then move to the correct distance for ascending)? Couldn't you also automate the ascent process rather than having to control it all manually with a remote? It seems like it's a fairly repetitive task that would be easy to set up macros for each of the different movements and their timings
In engineering this is what we call the "brute force" approach. Normally when somebody says "Just add more actuators!" you know a dumpster fire is in your future.
interesting project well for starters how about acceleromter(?) in the middle so it can calculate tip forwards and backwards for balance, simple switches in front on both side for step placement.
What about 3 vertically sliding links with each having a track so when the first link encounter the stair the track will pull it up while the rest will press it to the stair so that it has traction, and once its goes up the next one climb etc.
Thank you for your video, as I also need to make a 'scooter' that can go up stairs, and I am thinking of using a segway type device that has large diameter wheels and you can shift your weight forward. It needs to go faster than you can do it yourself with your legs or it won't be useful....It would be great if you could figure this out. The track system might work if it has large diameter more like a wheel. Cheers, Jack btw, a simpler method is to fire a hooked cable to the top of the stairs and just winch yourself up, doesn't need anything but a single winch and some wheels. Can go back down too. And very quickly too if you don't mind some bouncing.
Such a cool concept! Well executed too. I'm curious if you think it can make it down the stairs in its current form? Also a stair-lift robot that can interface with other robots would be amazing. Retrieval / delivery robots on each floor plus the stair lift robot to take a mass from one level to another. What an awesome home-automation system you'll have soon!
I was thinking you could mount a rail To the wall just have it extended out and enter so and have the robot be able to connect to that and drive up the rail… who knows if that would work though. And in theory that should be able to make it down the stairs it just goes in reverse
Granted I'm only half way through - but the horizontal expand function seems like it might make wheels redundant - i.e. inchworm style movement up the stairs rather than rolling?
Make the weight mechanism have 3 stages, when you lift the first wheels the system automatically first has to have the weight on the first stage, when you lift the middle the weight automatically moves to middle, and when you lift the last wheels the system automatically moves the weight to the last stage, this removes two buttons and stream lines the process so that if further autonomy with sensors may be easier.
I'd like to see a Treadwell type droid go up and down stairs. Using the arms to help start going up. Then tilting the tread bed at the base of the body so the body remains upright.
Interesting concept, but if I want something, I just going down the stairs to get it myself. Now if you were able to get a robo flying drone that can carry a load, that could bring things up fast.
Love the idea with the sliding mass! But doesn't this make the back section with the omniwheels redundant, because with the sliding mass you can stop the front sections from tumbling down
@@jamesbruton why not? In my mind, to go onto the step above, the middle lifts the front up by turning the lead screw, and then the whole assembly moves forward. If there were only two sections with a moving counterweight, I think the middle could be picked up: when the front is on the step above (at a higher position relative to the middle because of the lead screw turning), wouldn't making the lead screw the other way (ie in the direction if the front were to be lowered) pick up the middle?
Join my DISCORD! discord.gg/fc6MedG7eW
INSTAGRAM: instagram.com/xrobotsuk
TWITTER: twitter.com/xrobotsuk
MASTODON: fosstodon.org/@XRobotsUK
You could make the weight shift automatically depending on which linear motor you enable? I.e. moves forward before moving rear wheels etc.
Camera bot doing a great job holding the camera!
Got a little seasick at the end, maybe need a bit more damping.
It does a good job, however needs more damping indeed. Switching point of interest instead of accurately following it is what's needed.
And personally I'm used to a signature steady wide open shot with a lot of machinery in the frame. Change is not always good.
A lot of videos today are dynamic and shot from a handheld cam. James style stands apart and that is good IMO.
@@danielf3623 Same, at the end it was a bit too much haha
@@kittguy And another bunch of videos are filmed on a tripod, with wide shots. James' style wasn't special at all, in fact a robot filming him IS a special style he needs. Just needs some tuning, that's all.
@@daftbence Solid point actually, I haven't thought of it this way. 👍
It's a VERY good machine to carry a quite massive load and for keeping it level and stable. But for it carry you a beer while you're feeling lazy... well I think it would be warm by the time it got there lol. You did mention that it was overly complicated, and I agree... for SIMPLE tasks. But it really is the perfect design for what it was based on (cleaning the stairs)
brits drink their beer warm anyway
If I were to do this project, I would cheat and mount a track on the wall that the robot could attach to and drive itself up the wall like one of those stair-lift-chair things.
Or attach and detach itself to an installed winch or crane.
That would only help for a home use that's planned ahead of time like that - not really as robust as a solution as this robot that could go up many different types of stairs.
EDIT 12:32 in the video he has the same idea haha I wrote this comment before finishing the vid
He's English, warm beer is the go.
If anything, there's a risk he'll have to drink his ale cold.
instead of 3d printed part for ads it was more struddy to use only aluminum parts and make a mini fridge as counter mass
That worked so much better than I was anticipating. Excellent results
Brilliant!
As a proof of concept you've created a lift for elderly or handicapped that doesn't need rails and is portable. Truly brilliant!
The new camera rig seems like it's working out great, but with this project, because you've already added the 2kg weight shifting element, why not just use two sets of balancing wheels and a single actuator between them?
Funny - I had the opposite thought - since it's using 6 wheels, why have a shifting mass when you could just add a static mass to the middle section? :P
i think a servo with a stability plate at each set of wheels would assist in this if needed
CAMERABOT is doing a great job man. i think you're the only one i know off on youtube to have that. brilliant
Gotta love the camera work 👍
Besides being educational, your videos are always fun to watch. The music for this one was perfect.
Amazing. I’d love to see an autonomous version of this that can go up and down. But also excited to see the Roomba stairlift!
So very cool, well done James and thanks for what you do for the Robotics community and STEM learning in general.
Instead of sensor, limit switch for hitting side of the stairs is probably easier to implement.
Maybe some problem though with that fuzzy carpet
@@bluemamba5317 Maybe a limit switch with a bigger bumper?
I'm pretty sure sensors were limit switches when I was learning robotics back in the 80s. There were infra-red sensors too, but there's nothing wrong with an, *ahem* _simple touch sensor_ AKA limit switch. :)
Onshape is about 3x the prices of FUSION 360. Why would I switch?
Excellent project 👍
Love the multiple different types of pullies and different engineering designs in this. On a quick note you should try to use mecanum wheels, they give lots of control when used right and you can go in almost any direction you want!
What would be super cool is if someone could make like a rail system around a house that bots can attach to and ride around. For the stairs the rail could go up a wall and then across the ceiling out the way.
This is what I imagine a stair climbing Dalek would look like. Pretty interesting design though.
I gotta say I switched to Onshape about a year and a half ago when I switched to Linux and coming from fusion I gotta say I love it and I love to see them sponsoring you
not going to lie but I use Onshape and it is pretty good for most applications including 3D printing
A few weeks ago I was thinking about the same idea, designing a segmented stair climbing robot. Nice to see someone doing it.
Could you make the counter weight automatically balance the robot to remove a few steps from the climbing sequence?
nice automatism . You can also try triangular wheels (one wheel on each vertices), all three powered by a central sprocket , and the triangle turns and grips . Like a powered climbing stairs cart
Love the robot camera Is doing a great job
Yeees! The stair roomba is coming! Thank you!
You make it look so easy. You must have a ton of experience. Wish I could learn from you in person
New camera guy is really putting in the work
I love your soundtrack!
robot cameraman john is legitimately really good. its a good handcam feel. John has been replaced!
Impressive! James does in weeks what industry does in many months and universities take years to accomplish.
🤣 You think a university would only take "years"? 😂😂 An academic would sit on a problem like for his/her/its entire career. And in that time he/she/it would build a medium size research group, obtain squillions in taxpayer funded grants to fund said group, reach full professor relatively early compared to a more esoteric academic colleague and then spend the rest of his/her/its career dining out on his/her/its one pony trick without any of his/her/its sponsors ever really achieving a commercially viable product. Even more sadly he/she/it will successfully supervise higher degrees and they too will follow in his/her/its path at his/her/its institue or some other. And so and so on and so on. 🤣😂🤣
Hey Jim, I'm still loving the autonomous camera man 😉
Quite the accomplishment James.
It's not the robotics I'm impressed by, it's the time to finish that is amazing.
i appreicate the background music during the build montage!
Aargh! he always finishes projects that I am working on before I do
Fantastic, thank you!
This is an awesome build...
Your robot camera man is still killing it. Good job James, you're the best evil scientist on UA-cam.
Thanks, although opinions are mixed ;-)
Loved the new vid, cam bot is working great! I miss the old 3d print music though.:(
Objective achieved. Nice.
Bold choice on the new music!
Fantastic as always. I thought that seven minutes to get up the stairs was quite impressive. :)
A few weeks back I thought that your motion camera rig was another amazing project. However, when you use it now, there are parts of the video (like the closing section here) where I have to look away as they make me fill quite ill very quickly -- the mimicking of a drifting hand-held camera with no stabilization is too good! :) But this may be just me.
you are underrated engineer. woow I m bafed
I love it.... Glad were back to awesome robots..
Cool video. Looks like you got the robot camera rig working. Nice! Mahalo for sharing! : )
Really good explanation of the problem and your solution 👍
Your builds were already impressive with the BB8 but this is very cool.
Honestly thought James had hired a camera man for a bit. Then I remembered the last video! Great work James!
If the compartment of the battery can move to accommodate for the change of weight in the direction needed, that's a huge win for efficiency.
Now imagine being able to do this with more than one battery separately for even more control and capacity.
Really cool video and what an amazing 3D printers ! Mouth watering...
I know for a FACT you used your camera bot and damnnnnnnn that lil homie was build by a pro cause it didn’t miss a shot. Love the video have a great day!!
I see you use your camera robot for filming now. Thats pretty awesome that you get something really useful out of your inventions as well from time to time.
Love the dynamic camera!! Also, your content is pretty dope
I keep forgetting you're using the robot camera operator. Awesome.
7:10 Why not use the battery pack as a counter weight or mass, instead of adding even more weight/mass?
Awesome job man. One suggestion. Instead of shifting weight, you can use flywheels. You can control their speed and direction to adjust the angular momentum. I think it will be comparatively simple and compact.
In order to automate, can you simply detect the first stair, input the length and height of each tred and then let it run?
the camera work is pretty spicy in this video, very nice
That's the camera robot?
Great video! I had a thought though, and I'm not sure if I'm missing something (in which case, please correct me), but if you have the moving mass that can change where it balances, why couldn't it be just two segments? Like, it puts one segment up one step, shifts weight onto that one, lifts the back segment up, transfers weight onto that one, etc.
Really a great job. It is not easy to make a robot that can climb such stairs. Now maybe you can use it as a platform, add more automation etc.
Awsome! Great build!
I think it only needs to drive the front pair of wheels when the middle pair is at the same level as those front pair. This means you could simplify by having a simple gear or friction drive that takes the drive off the middle pair and supplies it to the front pair when they are at the same level?
a couple rails with a section of teeth in the middle would make it easier for a small robot to quickly climb the stairs. depending on size you could get it to corner on the rail and even a weak, slow climb would still probably be faster than all the stop and go with this bot. this robot is amazing in its own way and if it was only a step or two it would be fine but 7 min is a while and then you need a good bit of sensors to keep it from just driving off the edge.
Imagine getting your morning coffee from this.
At least, I think you can automate the stair climb with 1 sonar sensor on the front:
You still have control of the driving of course. You can drive it up to a step, and hit a button that will start the stair climbing (maybe with the same button to stop the auto-climb so you can take over too).
The program will first make sure it's close enough to the step, maybe even bumping into the step.
It will then raise the front section until the sensor doesn't detect a close object.
It then drives forward a set distance, stopping if the sensor gets too close to the next step (if there is one).
I'm not going to write out the whole thing, but I think you can get the idea for the algo from this. Just one sensor needed! :)
If you can find it have a look at the British Army's "wheelbarrow" from the 1980s. It is an EOD robot that carries tools and can negotiate stairs no problem. It uses tracks and has a hull that slides fore and aft for centre of gravity control. Well done on this design...
For a "proof of concept" or version 0 it's amazing!
Not bad for a first step... I'll see myself out now.
In all seriousness though, it reminds me of the way a baby climbs stairs.
Functional, effective, but very slow.
Very interested to see what else comes out.
I would love to see more Robot Dog stuff, I think you can keep pushing it
How about a hopping / pogo stick robot that can hop up the steps?
Would need some kind of active damping on the beer carrying arm though.
Sounds right up your alley with some reaction wheals and or gyro's for stability and inducing forward motion and rotation etc
Wouldn't you simply drink the beer before climbing the stairs? And what is so important at the top of the stairs anyway? Tell her to come down, you're too busy watching the cricket
Always enjoy seeing your solution to hese problems. Adding a payload area seems like it might get dicey. You're already fighting balance issues. Seems like figuring out additional weight distribution could be a challenge. Thanks for sharing!
But can you make stairs that climb robots?
Next joke?
Excelente work... Greetings from colombia.🎉
Rather than sensors for detecting the distance to the next step couldn't you use switches on each side (front left and front right) and let the robot bump into the stair like the early roombas did to detect the stair and then orient itself correctly (straighten out and then move to the correct distance for ascending)?
Couldn't you also automate the ascent process rather than having to control it all manually with a remote? It seems like it's a fairly repetitive task that would be easy to set up macros for each of the different movements and their timings
Great ! You should add this mass moving system on your useful robot so he can actually lift heavier objects without balancing problems.
Cool, but I wonder if you could build something like the wheelchair tank style treads in the Mark Rober video .
This is an AMAZING project! Can the robot go downstairs? I think it can, just reverse the steps should work
Nicely made as always! I think with mechanical switches tho, you could double the motor use, rather than simply using the main motors one at a time
One of these days we're going to see you make a 3D printed full mech/Gundam suit.
In engineering this is what we call the "brute force" approach. Normally when somebody says "Just add more actuators!" you know a dumpster fire is in your future.
interesting project
well for starters how about acceleromter(?) in the middle so it can calculate tip forwards and backwards for balance, simple switches in front on both side for step placement.
What about 3 vertically sliding links with each having a track so when the first link encounter the stair the track will pull it up while the rest will press it to the stair so that it has traction, and once its goes up the next one climb etc.
Hi Jim! Love your creations! The plastic they're made out of remind me of my 90s childhood for some reason. 1:54am in NY here, cheers!
ultrasound worth a try for the automation problem? you could have lots of transducers everywhere measuring everything. it might go really fast
smart design !
Thank you for your video, as I also need to make a 'scooter' that can go up stairs, and I am thinking of using a segway type device that has large diameter wheels and you can shift your weight forward. It needs to go faster than you can do it yourself with your legs or it won't be useful....It would be great if you could figure this out. The track system might work if it has large diameter more like a wheel. Cheers, Jack
btw, a simpler method is to fire a hooked cable to the top of the stairs and just winch yourself up, doesn't need anything but a single winch and some wheels. Can go back down too. And very quickly too if you don't mind some bouncing.
enitre video i was thinking of doing this to a rumba, and then you said it.... can't wait for that one... :)
Such a cool concept! Well executed too. I'm curious if you think it can make it down the stairs in its current form?
Also a stair-lift robot that can interface with other robots would be amazing. Retrieval / delivery robots on each floor plus the stair lift robot to take a mass from one level to another. What an awesome home-automation system you'll have soon!
I was thinking you could mount a rail To the wall just have it extended out and enter so and have the robot be able to connect to that and drive up the rail… who knows if that would work though. And in theory that should be able to make it down the stairs it just goes in reverse
Would there be a room in the robot to mount a cleaning equipment (like vacuum or detergents spray with towel) ?
I look forward to your post amazing job again. Thanks
Granted I'm only half way through - but the horizontal expand function seems like it might make wheels redundant - i.e. inchworm style movement up the stairs rather than rolling?
Make the weight mechanism have 3 stages, when you lift the first wheels the system automatically first has to have the weight on the first stage, when you lift the middle the weight automatically moves to middle, and when you lift the last wheels the system automatically moves the weight to the last stage, this removes two buttons and stream lines the process so that if further autonomy with sensors may be easier.
Cool copy.
Is it possible to make this prototype move 10 times faster.
Thank you very much for your endeavor.
Is the axial load of the ball screws entirely held by the spider coupling set screw??
I'd like to see a Treadwell type droid go up and down stairs. Using the arms to help start going up. Then tilting the tread bed at the base of the body so the body remains upright.
Interesting concept, but if I want something, I just going down the stairs to get it myself. Now if you were able to get a robo flying drone that can carry a load, that could bring things up fast.
Bring back the old printing moment music :'( (awsome as always)
Cool build but guess the beer will not be cold once it arrives 😉 Thanks for sharing 👍🏼
It might take a while, but you'll really enjoy that Pepsi from 1982
this is very cool and i love the solution but i think it needs to go FASTER!!! i hope next episode is a faster version. you make cool stuff 😊
Love the idea with the sliding mass! But doesn't this make the back section with the omniwheels redundant, because with the sliding mass you can stop the front sections from tumbling down
You can't pick the middle up...
@@jamesbruton why not? In my mind, to go onto the step above, the middle lifts the front up by turning the lead screw, and then the whole assembly moves forward. If there were only two sections with a moving counterweight, I think the middle could be picked up: when the front is on the step above (at a higher position relative to the middle because of the lead screw turning), wouldn't making the lead screw the other way (ie in the direction if the front were to be lowered) pick up the middle?