new web: http://bdml.stanford.edu/pmwiki
TWiki > RisePrivate Web>PrivateRiSE > NewProposals>FootActuation (10 May 2006, MarkCutkosky)

Foot Actuation

New stuff:

Movies showing new prototype:

Other comments:

There are a few changes I will make to the design as posted previously. Here is a brief list of changes that will be made. Updated SolidWorks files will be posted soon.

  • Change cam angle so it has a longer shallow-slope part so it can wedge foot down better, also make the steep part shorter and steeper.

  • Make bump in FootAttachment piece (piece on the hinge that goes up and down) bigger to improve cam wedging action

  • Cam needs groove all the way around to permit full rotation

  • Cam should occupy more of a full circle (probably 330 degrees) rather than the ~240 degrees it is now

  • Might have to make changes in how the piece that rotates up and down (FootAttachment in the solidworks files) is attached to its hinge

  • Put screw hole near the running foot that holds cam cover down (currently it is primarily held by the two screws on one side)

  • Make cam cover go all the way down to the flat plate on the bottom, rather than having a ~3mm gap as it currently has

  • Strengthen bottom plate by adding a ridge along the length of it (on the bottom)

  • Check height of screw holes in hinge, and also check tolerances on cam and cam cover. How much friction will aluminum have?

  • Increase spring stiffness and include a preload.

-- AlanAsbeck - 24 Feb 2006

Previous stuff:

The general idea is to have a servo mounted on the main body of RiSE, and have cables go down to the ends of the lower legs. On each lower leg will be mounted a cam which will control a hinge that will push a spiny/sticky foot up and down so either it or the dactyl can be used. Because of the non-backdrivable cam design, the foot can be moved to intermediate positions between "up" and "down". The system will be a configuration-changing mechanism and will not be used to, say, disengage a foot on every step. Also, our current design includes a running foot mounted on the end of the lower leg. It has been designed so that the polyurethane is quite thick and won't crack as happened with the previous running feet.

Nomenclature: Please call this a "Foot Actuator" and NOT an "Ankle". The Ankle is the part of the foot that contains the ball joint and permits yaw motion. This device is a shape-changer and will act approximately like a rigid link.

Finally, for those who are concerned about code development not being completed in time: in the worst case, we can attach a remote-control car receiver to RiSE with a small battery, and control the Foot Actuators by hand using a remote-control device. We've done this a lot in testing SpinyBot and other early Spiny/Stickybots.

-- AlanAsbeck - 21 Feb 2006

Pictures of a crude prototype with cam and foot and dactyl.

Dimensions of cam in prototype are the same as in SolidWorks model (shown below). Please ignore the metal threaded inserts on the cam, they will not be there in a real version.

  • Prototype - cam engaged, foot down:
    img_camdown.jpg

  • Prototype - foot up, back view:
    img_footupbackview.jpg

  • Prototype - view from end of foot showing dactyl:
    img_footupviewfromend.jpg

Pictures of SolidWorks model

The SolidWorks files can be downloaded here:

  • Overview of Foot Actuation idea:
    topviewfromheadcopy.jpg

  • Overview showing spring and foot:
    topviewfromheadcopy2.jpg

  • Top view looking from back of foot (looking from tail):
    topviewfromtail.jpg

  • Bottom view:
    bottomview.jpg

  • Cam pushes foot down:
    campushesfootdown.jpg

  • Cam is retracted, foot in the "up" position:
    camretractedfootup.jpg

  • Transparent cam cover showing cam and slot for cable:
    transparentcamcovercopy.jpg

  • Side view looking from the tail (back of foot):
    sideviewdactyl.jpg

  • Side view showing where dactyl goes: Note that in this drawing the dactyl is smaller than it will be in real life (see pictures at top of this web page)
    sideviewdactylcopy.jpg

Pictures of servos with cable drive, currently on Main.StickyBot

The cable attachment on the Foot Actuator will be very similar to the mechanism shown in the pictures below, with the cable going between the cam and the cam cover in a little groove/slot. The cable will come out of the cam cover and go right into a sheath, a little tube that is glued to the cam cover.

In addition, the servos on the main body of RiSE will have cable drives like these.

  • Servo with cable drive attachment:
    img_servocabledrive.jpg

  • Another servo with cable drive attachment:
    img_servocabledrive2.jpg

The cable sheath will have a stretchy section at some point, allowing the servo to drive the cable to, say, two positions, the 'up' and 'down' positions. Then, the cam will go as far as it can towards those positions, and the stretchy cable sheath will accomodate any difference between the actual cam position and the position the servo is driving it to. This is useful so you can drive the cam into the 'down' position as far as it can possibly go until the foot hinge becomes wedged between the cam and hard stop on the bottom.

 
This site is powered by the TWiki collaboration platformCopyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback