I have built a LEGO swerve drive, a Shift-On-The-Fly LEGO gearbox, and a LEGO hexapod… But now it is time to move on the bigger and better things. Thanks to Tytus Gerrrish’s suggestion in this thread I am now working on my most difficult challenge to date. Unlike my previous projects where I just show you all the final project I thought I would post updates as I work. I am not doing this so much for you guys but more for me. I want to see what people are thinking as I work on it, and possibly get some group opinions/suggestions along the way.
LEGO Biped Theory of Operation
The type of biped I am attempting to build is a COG (Center of Gravity)-shifting type. This means that the weight of the robot will be shifted over one foot of the robot, allowing the other leg to be lifted and swung forward to take a step.
Walking
As stated above to walk the robot will put the weight of itself over one foot allowing the other to be swung forward (or backwards). It will then lower its foot and repeat the same operation with its other foot.
Turning
In order for the robot to make a right of left turn it will use the same operation it used for walking forwards and backwards with one slight difference. After one foot has completed a full step, the second foot will only complete a half (or less) step. Because of this one side of the robot will be traveling farter then the other side, causing it to turn.
Prototype Pictures
So far I have only completed a prototype of one foot assembly.
The first picture is of the foot in its normal upright standing position. As you can see the foot is just a large square shape. This may or may not change depending on any balance needs later on.
In the center you see a LEGO gear motor with an 8 tooth gear on its output shaft. It is meshed with a 40 tooth gear that is attached to the foot assembly. This allows the leg to tilt as shown in this picture. This is what will be used to shift the weight onto one foot.
On the top of the leg you see a large turntable. This is what will be used to swing the legs. The gearing at the top meshes with the gear on the inside of the turntable, allowing it to be powered by a motor. This I idea was borrowed from another builder who uses a “powered turntable” to drive a wheel on one of his creations.
That is all for now. Any comments/suggestions/complaints are welcome!
Tytus: I thought you were serious… well either way it is still a great project that I am having fun with.
even so, i still dont see any legs
Those two pictures show basically all there is to the leg. I hope you were not thinking that by leg I meant foot, ankle, knee, ect… I would neither have to motors/parts for all the joints, not to mention the computing power to create something that walks like a human (controlled falling).
As to shifting the weight, I’d say having a moving mass would be easier than having ankles that turned. If you geared down a motor to an off-center weight that rotated once every two steps (like one rotation every time the left foot took a step) you could have a large percentage of the weight of the robot shift from one side to the other for each step and thus keep it from falling over.
This may be more difficult than what you already have set up, but it’s what I plan on doing whenever I get around to making a biped.
If you are really serious about building a bipedal walking machine, let me make two suggestions:
read everything you can about dynamic stability and dynamic locomotion. Bipedal walking is not a statically stable process. When you are walking, you are constantly “almost falling” and using the moving foot to arrest the fall. A lot of people wasted a lot of time trying to make walking machines that were completely stable at any moment in time. It was only after they realized that walking is an inherently unstable process, and the that stability is due to the constant motion of the center of mass, that real progress was made. To avoid a lot of blind alleys, think about how you will keep the center of mass of your entire machine moving constantly - don’t think that it has to be balanced at any given moment.
find and read everything you can by Marc Raibert. Marc founded the MIT Leg Lab, before he went off to Boston Dynamics, and is one of the world’s leading experts in bipedal legged locomotion. Everyone else is just now catching up on the work he did 20 years ago.
find and read everything you can by Marc Raibert. Marc founded the MIT Leg Lab, before he went off to Boston Dynamics, and is one of the world’s leading experts in bipedal legged locomotion. Everyone else is just now catching up on the work he did 20 years ago.
Just out of curiousity…are you familiar with Steve Hassenplug (Member of the Lafayette Lego Robotics Club, featured on slashdot and tech TV, one of the advisors on the lego league teams i am also an advisor on)
Check the backgrounds of the Asimo team. See how much of their work traces back to Raibert’s early research at MIT, and prior work at CMU. The big difference is that Asmio basically records human walking and then plays back the motions, while Raibert’s work is based on actually understanding how walking “works” and creating an algorithmic solution to the process of walking. Raibert’s robots were running (and skipping and hopping) ten years ago, while Asimo can barely manage a slow walk even now.
From what I have read, Dave is very correct (as he usually is).
It is interesting to see how robots have developed over the years. An excellent book to read on this subject is Flesh and Machines , written by Rodney Brooks of MIT.
This book is educational and inspirational, while even being easy to comprehend.
Put it on your Christmas list and hope that you have been good this year.
Pauluffel-
My first idea when I was thinking of making the biped was very similar to yours. I was thinking about making a track from the rack gears across the top of the robot. Then I was going to have the RCX on a car with gears for wheels to drive left and right on the track to shift the weight over each leg. It was going to be very similar to how the tram works on my copy machine.
Dlavevy-
Thanks for the tips but im not to sure that the RCX will be able to read the sensors fast enough/ control the motors fast enough to perform any controlled falling.
Also thanks for the link, there is a lot of great information on that site.
I also just wanted to share this link that I found with everyone. It has a bunch of videos of different bipedal robots.
The first set of robots (Sony SDR-4X) sing a little song and dance. Also check out the robot that walks backwards as the man pushes it with his finger. I don’t know why but that one creeps me out.
After working on the biped for quite some time now, and running into a wall, I figured it was time for an update.
I have built a second identical leg and connected the two with a simple body. You can see an overall picture of it here. In this second picture you see the robot as it balances on one foot. The problem is that the motors in the foot do not have the power to lift the robot into that position. I have come up with two possible solutions for this.
The first would be to simply redesign the feet to incorporate a lower gear ratio.
The second idea is to not use motors at all. Instead I would replace them and use pneumatic pistons to tilt the robot over one foot.
Again any comments/suggestions/complaints are welcome!
Originally posted by rlowerr_1 *
**
Thanks for the tips but im not to sure that the RCX will be able to read the sensors fast enough/ control the motors fast enough to perform any controlled falling.*
Actually, the RCX can probably keep up with the overhead of a basic dynamic stability control system. For example, just one RCX can be used to control a simple Segway build from Lego parts.
You will want to be careful about monitoring the cycle time on the control loops. If the time that it takes to complete one control loop begins to approach either the sensor refresh rate or the response time of the motors/actuators, then you will need to back off a little. But you can do this by either decreasing the sensor processing required (tough to do), or by just adding an additional RCX and sharing the processing load between multiple units.
Alternately, you could break the entire system into similar modules and assign one RCX to each (e.g. each leg is a module - just mirror images of each other - and the torso is just a big leg module, etc., each with it’s own RCX). Each RCX is then programmed with the control loops (and slightly higher-order behaviors) required to just run that one module. The fun part is figuring out how to have the RCX units pass information between themselves to create coordinated motion. This quickly gets into a really interesting problem (this is where Rod Brooks/MIT started with his subsumption architectures in the mid-1980s - the same Rod Brooks that Andy mentioned above).
As I noted above, if you really want to get serious about this there is a lot of information to explore and a lot of good work that has been done that can help you get started. But the best part is that walking machines are just plain cool! - if you can pull this off, you will have a lot of fun and the undying respect of us all!
Ok it’s time for another update after working on this thing early in the morning until late at night for the last few days.
After much messing around with trying to fit two motors/lower gear ratios in the leg assemblies I gave up and decided to give pneumatics a shot. This is the first time I have used pneumatics in one of my projects, but the result I have received from them is quite pleasing.
From this shot you can see all the changes. The first most notable one is that her legs got a lot longer. This was done to allow the pistons to be mounted and to give more room in the leg for the pneumatic pump/switch gearbox. Also you see that the feet have taken on a wider shape for more balance when the robot shifts its weight.
Looking at this picture shows you what used to be a major problem, or what I though was a major problem. When the robot tries to go from standing position (both feet on ground) to balancing position (balancing over one foot) instead of lifting the robot over one foot, the robot gets on its “tip toes”. At first I thought the project was toast, but after a break from LEGOS I came back and quickly realized that when one leg was on its “tip toes” it shifted the weight of the RCX just a little bit over the other leg, allowing that leg to tilt the robot over itself.
You can see the robot balancing on one leg in this shot. Its hard to tell in the picture but the left foot is about an inch off the ground.
All that is left to due is make two motor mounds to drive the pump/switch assemblies and it should be ready to take a test step!
This picture is of the biped taking its first non human powered step. It is not being controlled by the RCX, but instead by the battery box in the left bottom corner of the picture. I am very close to having this thing walking on its own but there are just two problems that have to be delt with first. The first minor problem is that the motor that controls the spinning of the body around the foot that it is balancing on is not quite strong enough. I already have plans to gear a second motor into the system to help it out.
The second major problem is that one of my small pneumatic pumps does not seem to pump. I am not sure if this is something I can fix myself by opening it up, or if I should order/ask for another one from LEGO. But until I fix it/order a new one this robot aint walking anywhere!
The last thing I wanted to talk about is a name for this robot be sides the LEGO biped. Anyone have any ideas?
There is a link to a lego biped that I built as well as another I built with RC servos (Mechadon) using the multi-processor approach that Dave suggested. On the lego 'bot I used a single motor to move both ankles and shift the weight. As you found out, getting a very low gear ratio is key when using motors. I also found that using a worm gear drive at each ankle helped a lot because it couldn’t be back-driven so the motor didn’t have to exert torque continuously to hold a position.
One of my other robots, MiniMechadon, will use the “tip toe” technique to begin shifting it’s weight similar to your 'bot.
Looks like you are on the right track and have made great progress so far!
