Group Actions
Engagement
  • Author Author: dubbie
  • Date Created: Date Created
  • Views 953 views
  • Likes 6 likes
  • Comments 5 comments
Related
Recommended

Walking Robot #1 : Collecting Some Ideas

dubbie

For some time I have been interested in creating my own walking robot. Not so much the two legged human like but more replacing wheels with something resembling a leg or leg type motion. I have sort of designed and made a four legged quadruped type spider thingy which was OK but was difficult to get walking, as in, I didn't get it to walk. I could never work out the gait. It is all to do with the centre of gravity as with a leg at each corner if you lift one leg the whole thing falls over. So you have to shift the centre of gravity to a more stable three legged point and then lift the fourth leg. Plus it needed two micro servos for each leg. For this type of spider type walking you either need six legs (which requires 12 servos) or three servos per leg (which also requires 12 servos). Once into the high number of servos region you then run into power supply problems, which I solved by using two battery pack, but this all makes it heavy, expensive and difficult.

 

I thought there must be a better way, so I have obtained an Elegoo Penguin robot. This has two servos per leg and uses big wide feet to solve the balancing problem, although it will still fall over. It does walk albeit quite slowly and really only on very smooth surfaces. See video below:

 

 

This is a possible approach and I think it has some good pointers. Particularly the use of big feet to help with the balancing problem. It maybe that if some sort of knee joint was added this might walk a bit better.

 

An alternative approach is provided by a toy called Tobbie which uses some sort of six legged spider motion. Interestingly all the batteries, controller and sensors are in the top half of the robot which enables it to rotate through 360 degrees without getting wires trapped inside. It does also make it top heavy and the centre of gravity is much higher but a wider spread of legs is used to overcome this. The video below shows the top half of the construction, with the long pin showing the movement provided to the bottom six legs, which is some sort of up/down rotation. It looks like it should be something that could be reproduced with some 3D printed parts such as cams and levers. Something to look more closely at I think. Apologies if you cannot hear the audio in the second half of this video but the Tobbie mechanism makes a great deal of noise and tends to drown out all the other audio.

 

 

Well, I'm off to finish making Tobbie and to think about how to make my own 3D printed walking mechanism/robot.

 

Dubbie

  • in reply to genebren

    Gene,

     

    The Penguin does have plastic pads to stick to the bottom of the feet, which do reduce the slip. I'm not entirely such they helped as it seemed to work better without them, but then again the head of a small bolt in the bottom of the foot is slightly raised and could have been 'digging' into the surface. The sticky pad added is the same thickness as the screw head and creates a level surface.

     

    Dubbie

  • in reply to dubbie

    The larger feet did improve the balance, but they did not work as well on the turning.  With this simple walking approach a turn is accomplished by leaning to the side and moving on leg forward then leaning back onto both feet and finally with both feet in contact with the ground reversing the legs.  This action should create a drag/pivot, causing the robot to slightly pivot.  The problem that I had was that the feet would just slip without creating the pivot.  I suspect that weight was not sufficient to create any dragging effect.  The earlier version carried more weight (4 AAA batteries, 2 on each side, plus a heavier weight in plastic throughout), giving the feet a little more grip.  I tried adding various material pads (rubber/cork) on the bottom of the feet, but to no real effect.  The next step is print some bigger feet (I was using cardboard) and to add weights (lead or other material) to see if I can create enough grab to get Walky to pivot.

     

    I do really like the penguin ankle/hip approach, not just for it's smoother motion, but also because you can do so much more moves, almost dance like.

     

    Gene

  • in reply to genebren

    Gene,

     

    Yes, I did see your Walky biped robot which seems similar to the Penguin. It is a bit disappointing about the bigger feet not working for balancing, as I was thinking that might be my backstop if nothing else worked. Ah well.

     

    Dubbie

  • I recently worked on a bi-ped walking robot (Walky the Biped Robot - WalkyII the final chapter. ) for the project14 mover and shakers competition.  In it's early form, the gait was actual a little more stable (guessing that I magically got the weight and center of balance a bit better), but in the course of the re-design I had nothing but troubles with stability (constantly tipping over).  I attempted to compensate with much larger feet, but in the end it was never quite right.  I also was never able to get it to steer as well as I would have liked (maybe not enough weight to allow the feet to grip and pivot).

     

    I really like the penguin model and have thought about reworking the bottom half of Walky to incorporate a similar walking concept (I did decide to add two additional servo channels to the design of the controller to allow for 4 servos).  It will be interesting to see what approaches you try in your efforts.

     

    Gene

  • After you mentioning Tobbie last week, I saw it and had to click.  It seems so simple a method of control I'm looking forward to seeing what the leg mechanisms look like.

     

    I also like the old sounds it makes.  Reminds me of computer games in the 80's

     

    Dan.