Stickybot from inside MERL

StickyBot what's new? Microscope video of version 1 stalks attaching to a thin strip without normal preload. For the latest hierarchical directional materials see ClimbingAdhesion . For the latest platform see StickyBotIII . Key Ideas: Stickybot is an embodiment of our hypotheses about the requirements for mobility on vertical surfaces using dry adhesion. The main point is that we need controllable adhesion. The essential ingredients are: hierarchical compliance for conforming at centimeter, millimeter and micrometer scales, anisotropic dry adhesive materials and structures so that we can control adhesion by controlling shear, distributed active force control that works with compliance and anisotropy to achieve stability. These ideas are explained in more detail in RisePublications on Stickybot and Directional Adhesion.

Stickybot Links

• Stickybot movies
• Publications on Stickybot and related climbing robot technology
• StickyBotPrivate - Stickybot developments not ready for release
• StickyBotController - Development of gait + force control for Stickybot
• GaitPlanner - Matlab script that creates basic gaits
• AdhesionModels - dry adhesion principles, analyses, tests
• HierarchicalStructures - putting hair on stalks and over
• StickyBotMechanical - Mechanical design for StickyBot and SbTwo
• RiseSiteMap

What's new

July 2009: Stickybot III is under development, with the goals of providing more robustness and more controllable DOF per limb.

May 2009: S. Kim, et al., "Smooth Vertical Surface Climbing with Directional Adhesion" receives King-Sun Fu Memorial Best Transactions on Robotics Paper Award.

Feb 2009: StickyBot (and Sanjay) accompany Bob Full at TED 2009

October 2008: Using new 2-layer hierarchical adhesives, Stickybot can climb wood panels, painted metal and other surfaces without changing control settings from those used on surfaces like glass.

• Details in Asbeck et al. 2009 presented at ICRA09
• (Short video)

September 2008: StickyBot featured on Science Channel.

March 2008: Stickybot and other bio-inspired creations featured in the April 2008 issue of National Geographic magazine.

Aug 2007: Videos of Stickybot and two-toed version of SB2 climbing outside, rehearsing for DarpaTech 2007

July 2007: We finally are starting to get some hierarchical directional, patches that look promising... Details in (private) HierarchicalFabrication? page. (To be made public soon -- meanwhile you can look at HierarchicalStructures general concepts page.)

25 May 2007: SbTwo is climbing, almost as well as StickyBot

14 April 2007: New papers and a new short video presented at IEEE ICRA07 (5 MByte, requires Quicktime).

29 March 2007: New foot pads - (they're blue!). Details soon, assuming they test as well or better than the originals. 05 Jan 2007: Sangbae Kim and StickyBot on PBS Wired Science 07 Nov 2006: Stickybot featured in Time Magazine, "Best Inventions 2006" 04 Sep 2006: Stickybot featured in Forbes Magazine as part of an article on "Seven Robots that will change your life." 29 Aug 2006: Directional adhesion explained in "Frictional Adhesion" paper with Kellar Autumn? in Journal of Experimental Biology. 3 Aug 2006: New sticky patches molded (very similar to the previous directional patches, with modified molding process). Testing is ongoing using a wall that contains a 6-axis force platform to measure the differences in contact forces for directional versus non-directional adhesive patches. We should see much lower preload and detachment forces with the directional patches...

Barrett instructs Stickybot to climb over the force plate

20 July 06 -- Video of climbing with diagonal stride gait (65MB). Ongoing work is focused on force sensing & control and various other refinements.

28 June 06 -- (Private) SbTwo Video of climbing with diagonal stride(6MB) (versus tripedal) gait at 0.8m/min (faster). Not only does it allow faster climbing but also it reduces unwanted swinging motion that causes premature detachment.This should permit faster dynamic climbing, ultimately.

14 June 06 -- New patches to be sent away - (Private) PatchPrototypes

13 June 06 -- Movie of test results of anistropic patches (Private) AnisoPatches

05 June 06 -- To facilitate parallel work on anisotropic adhesive pads and the StickyBotController we decided to create a sibling for Stickybot: SbTwo (Stickybot 2).

Here is a first video of climbing with anisotropic adhesive pads (2 MB). --Sangbae Kim

Note that although the control is clearly not optimized it already climbs better and with lower detachment forces.

• More about SbTwo design modifications on its own new page.

24 May 06 Stickybot has experienced a small flurry of web traffic...

5 May 06 -- The following video clip illustrates anisotropic (directional) compliant adhesion, as compared to conventional double-sided tape. The anisotropic patches work on a range of smooth surfaces (glass, acrylic plastic, glazed ceramic tile). They stick when loaded in shear in the desired loading direction. Adhesion is very low when loaded in the contrary direction. This capability, inspired by the results from Prof. Kellar Autumn on gecko adhesion, allows a climbing machine to attach and detach the feet with minimum preload and mininimum disturbance. (This capability is also in contrast to the isotropic materials using microfibers, nanotubes, etc. that a number of researchers have been developing.)

• demonstration movie -- 45 MB version
• demonstration movie -- 2 MB version, lower resolution. --Sangbae Kim

20stock_45tip_380umhair.JPG

The material in the video consists of stalks at an angle of 20 degrees with respect vertical. The tips present a face angled at 45 deg. with respect to vertical.

17 Apr 06 --Spatular.JPG The First spatular tip at a (large) micro scale. These features are inspired by the spatular tips at the ends of gecko setae. They increase the real area of contact and increase adhesion. A new molding process, which involves a combination of capillary action and exposure of the top surface to atmospheric moisture, is used to created graded material properties so that the features are softest at the tips. A provisional patent has been filed on this and related methods for creating anisotropic adhesive materials. The next stage of development is to obtain similar "spatular" geometries at about 1/10 the size of these, with a correspondingly stiffer material.

01 Apr 06 -- Stickybot_040106.mov: Stickybot on MERL window (25MB Mpeg4). This short video shows Stickybot climbing a glass window and a plastic panel showing the attachment and detachment. In this video the robot is using isotropic sticky patches. Performance will be better with anisotropic patches. If you look carefully you can see that the attachment and detachment forces are higher than ideal.

30 Mar 06 -- large (50MB) PDF file of poster for SWRI April 4-5. The point is that we don't need more adhesion, we need controllable adhesion. The essential ingredients are:

• hierarchical compliance for conforming at centimeter, millimeter and micrometer scales,
• anisotropic dry adhesive materials and structures so that we can control adhesion by controlling shear,
• distributed active force control that works with compliance and anisotropy to achieve stability.

28 Mar 06 -- First climbing demo (80MB mp4). Overriding Mark's anxieties, Sangbae put Stickybot in an oven (40 deg C for 1/2 day) to accelerate curing of new "blue" toe sandwich pads (~1mm thick, with 300 micrometer cylinders). After recovering from a certain amount of oven-induced distortion, the linked video was produced. Sorry for large file size but you need to see the feet clearly. From the video it's clear that further improvements can be made in the gait and compliance tuning. Of course, directional friction pads and closed-loop control on the limbs (coming...) will help even more.

23 Mar 06 -- first successful operation in which Stickybot lifts and replaces 1 foot while keeping the remaining 3 attached on vertical glass, without external support. The glass is not particularly clean but there is extensive tweaking of trajectories for smoothness and load sharing. Smoothness is crucial. Otherwise the some of the stuck feet become detached while the lifted foot is moving.
Meanwhile, we are iterating furiously on multi-layer compliant quasi-adhesive pads. The current ones have a "sandwich" construction with upper and lower "skin" layers and a matrix of little cylinders in between.

• stickybot3-23-06.mov: stickybot 1st steps and pad design (24 MByte MPEG4 Quicktime clip)
• stickybotrunning.mov: running on treadmill

Previous Results

Stickybot lives, as of Winter 2006. As of March 5 2006 it moves, but it doesn't yet climb vertical glass surfaces (keeps popping itself off).

Moved the previous Stickybot toe work to GeckoFootApril05

Stickybot (mostly) assembled on the bench:

Nov 1, 2005: Stickybot in the rough -- parts as embedded in sacrificial support layer:

Stickybot parts and molds

Solid model

• StickybotImages.zip: Archive of Stickybot pics and microphotos of stalks