How to Make Your Own Robot Cheetah

by rptech in Circuits > Robots

13861 Views, 79 Favorites, 0 Comments

How to Make Your Own Robot Cheetah

20140402_094104.jpg
20140402_093801.jpg
20140402_112205.jpg

Hello Everyone,

This AWESOME project was created from scratch in a high school tech class. It was designed for a state-wide competition for a club called Technology Student Association. I designed and created the project with help from Joey Outar (for aesthetics/paperwork) and Edward Jacob Shayevich (for controlling). I would like to thank our advisers/teachers Mr. Robert Ziegler and Mr. Roy Tamargo for helping us create such great work of art --- I mean how many high school graduates can say "I built a robot cheetah!" While robot cheetahs by DARPA and MIT grad students were built on multi-million dollar grants, Fluffy the robotic cheetah was created within a budget of $250.

p.s. Not all slides are the same, later step slides have more pictures and videos

Materials

1) Servo Power Board from Servocity.com

2) 4.8 Volt 2000 milliamps Nimh battery

3) 4 High torque HSC- 430 BH Servos

4) Lego Sticks

5) Corrugated Plastic Board

6) 4 Single Stranded Copper Wires

7) FAO Schwartz Cheetah Skin 29 inches in length

8) Solder and Tin

9) Acrylic Brackets

10) 8 nuts and 8 Bolts

11) Foam crafted leg fasteners

12) Two 3 millimeter syringes

13) 3 Plastic Tubes

14) 3 Sticks of twist ties

15) One 180 degree Hitec servo

16) Velcro Fasteners

17) Plastic Grinder top

18) Duct tape

19) Lego landscape and ridged Lego stick

20) Luan wood ¼ inch (more easier to cut than plywood)

Research and Design

Capture.PNG

Before the designing phase started, tens of hours were spent looking at similar projects. Although projects by DARPA and MIT were a great reference, they were far to complicated to replicate or modify. Work by Alexander Spröwitz was quite helpful and helped set standards for Fluffy to meet. Limited resources and a tight budget crippled the imagination. As a result, the first design was very very very basic and was made for only one objective, to build a stable chassis. The image you see is of a 3D model I created in blender (an open source animation software).

Building the Chassis/Base

20140220_160937.jpg
20140220_160949.jpg
20140220_163517.jpg
20140226_155336.jpg
20140226_155340.jpg
20140226_155653.jpg
20140226_160911.jpg
20140226_160913.jpg
20140226_160935.jpg
20140226_160940.jpg
20140226_161418.jpg
20140226_161438.jpg
20140227_152213.jpg
20140227_152220.jpg
20140227_152348.jpg
20140227_152835.jpg
20140227_154212.jpg
20140227_154255.jpg
20140227_154257.jpg

At first, acrylic was considered for the chassis but a plastic was used because it was much lighter (it is also easier to cut slots into plastic than it is in acrylic). Rectangular plates were turned into oval plates so it would be easier to insert them in the stuffed animal. Four heavy duty servos were mounted on the plastic frame and fastened by small pieces of acrylic. To stop servos from wobbling or escaping the holster, two pieces of wooden blocks were placed on the side (which also prevented the small pieces of acrylic from wobbling) and a small piece of sand paper was place above the servos (to increase friction).

Creating the Legs

20140227_154234.jpg
20140227_163519.jpg
20140227_163531.jpg
20140227_163541.jpg
20140227_163549.jpg
20140304_160833.jpg
20140304_161130.jpg
20140304_161142.jpg
20140304_161149.jpg
20140304_161158.jpg
20140304_161251.jpg
20140305_152401.jpg

At first, the legs were intended to be balsa wood or acrylic but plastic sticks were easier to manage --- they can be joined together by pins and easily cut or modified. First set of legs were made of only one stick but to increase strength and stability, three short sticks were used for the legs. The stick in the middle was a bit longer so it can be attached to the servo top by single stranded copper wire. The wire was fed through the servo top and knotted, the sharp end was cut, bent, and duck-taped so the skin would not get stuck.

Primary Installations

20140305_161440.jpg
20140305_161756.jpg
20140305_161802.jpg
Robot Cheetah Leg Test 1

To test the movement and get a feel of how to control each leg, the servo controller, its batteries and a power switch was installed onto the chassis. Switch was installed by a slot whereas the controller and battery was joined to the plastic by Velcro. After that, we pretty much started playing with it.

LEGS!!!!!!

20140305_163428.jpg
20140305_163439.jpg
20140305_163527.jpg
20140305_164300.jpg
20140306_152012.jpg
20140306_152014.jpg
20140306_152738.jpg
20140306_152743.jpg
Robot Cheetah Leg Test 2
Robot Cheetah Leg Test 2.01
Robot Cheetah Leg Test 3
Robot Cheetah Phase 1

As you can tell, controlling the legs was the biggest problem. At first we thought, a flat base might help, then we thought a trapezoidal shape might help, then we moved onto small cylindrical holders. Styrofoam was used for these leg stabilizers because it was easy to cut and it could absorb the shocks when the Fluffy moved forward.

The Skin (Why It's Called Fluffy)

Robot Cheetah Leg Movement Within Skin
20140220_163411.jpg
20140306_154021.jpg
20140306_154025.jpg
20140306_154102.jpg
20140306_155013.jpg
20140306_155022.jpg
20140313_153445.jpg
10578989_680317665369606_409619100_n.jpg
10602724_680317728702933_2029320464_n.jpg
10609139_680317765369596_973649653_n.jpg

NO ANIMALS WERE HARMED IN THE MAKING OF THIS ROBOT

A stuffed animal was purchased from toys r us and was skinned by yours truly. The stuffing was removed while making a scary face (too bad I lost the picture) and the bean bag from it's arse was played with until it broke. Before installing the body, we kind of had some fun with the skin --- we would toss it across the hallway by its tail or just walk with it on our heads to freak out the other kids. You should have seen the other kids' faces.

Anyways, afterwards, the prototype was installed into the skin. The Styrofoam grippers had to be removed to make it fit within the skin. The paws had small bean bags in them to keep it steady and give it a flat leg. However, a cylindrical Styrofoam gripper was used later on so the legs would stay steady and connected with the paws.

The Electrical Components

10567363_680315862036453_1334124117_n.jpg
10602854_680315928703113_131952758_n.jpg
10615838_680317338702972_1999368804_n.jpg
10617301_680317168702989_1830598779_n.jpg
10617586_680316078703098_799897325_n.jpg

By inserting the chassis into the skin and attempting to make it move, we learned that sides of the oval chassis needed to be thinner, the top needed to be taller an the top/bottom needed to be shortened. We had outlined our prototype design for functionality rather than cosmetics. Whence we made the necessary changes and moved on to adding the rest of the electrical components.

The servos were linked to a 8 channel receiver which was operated by flysky TH9X transmitter. A servo power board was installed between the servos and a 7.2v 3000 mah battery. The wires from the other end of the powerboard were connected to the servo controller. In other words, the servos and the receiver were powered separately because they need different amounts of energy.

Head and Tail

10617682_369962299824538_1544427416_n.jpg
20140313_153818.jpg
20140313_153820.jpg
20140329_152248.jpg
20140329_152303.jpg
20140401_152144.jpg
20140402_093801.jpg
20140402_094247.jpg

Mr.Ziegler soon helped design the rack and pinion mechanism for the neck. A servo would move a gear which would move a rack and push a plastic syringe forward. The syringe would push water through a tube and move another syringe. The hydraulic actuator would exert just the right amount of push to make the head nod up and down. As for the tail, we pretty much stuck a plastic tube through the chassis into the tail and curved it half way.

Obstacles and Regrets

Major Obstacles & How We Overcame Them

1. None of us had previous experience in quadruped technology --- we researched

2. There were binding issues with the receiver and the transmitter --- rebooted the transmitter to factory default

3. The transmitter was controlling multiple servos at one time --- made a menu with personalized settings to calibrate new zero marks

4. The rectangular foam spacers did not work --- we turned them into trapezoidal prism and then cylindrical tubing

5. The head mechanism kept falling off --- made a housing for them

6. servos kept falling off of the chassis ---- increased friction with holster by covering the top with sand paper

7. The chassis as too big to stuff into the skin --- shaved off the extra plastic

8. The copper wire ties for the legs kept getting stuck in the skin --- covered them with duct tape

Regrets:

1. Because our club supervisor did not provide financial support, I had to buy most of the products (servo power board, two high torque servos, RC Transmitter+Receiver, batteries, etc...) myself --- our adviser Mr.Tamargo helped purchase the other two high torque servos.

2. Since it was tight budget and we were short on time, I did not get to install additional features like a GPS sensor on a collar, sound system to make it talk, or a mechanism to make the tail move.

3. It would have been great to make the chassis or the legs out of some sort of light weight metal -- I tried making linkages (you can see it in one of the pictures for step 3) to make leg joints but the plastic was too lose to install a spring mechanism. If we had a laser cutter, we would have made a stronger frame or more flexible leg mechanism.

Ta-Da!

20140402_112205.jpg
20140402_112246.jpg
20140402_112249.jpg
20140402_112302.jpg
20140402_112305.jpg
20140402_112322.jpg
10543795_680316178703088_848721126_n.jpg
10569192_680316302036409_312985420_n.jpg
10617301_680317168702989_1830598779_n.jpg

Thank you for looking at our project, hope you liked it. Here are some pictures of the final version of the inside part. Few more pictures and videos were taken by others; I am in the process of collecting them (will soon post more pics). If you liked it, please vote. Please feel free to post any questions or comments.

p.p.s. I hope you saw that we didn't want to just copy someone else's designs or knew everything that we were going to do. For the most part, there were multiple phases of learning while creating. There were so many factors that played into making this beautiful abomination. We came together to overcome many obstacles and I think that made this project special.