SpaceSpeak is proud to announce a forthcoming FeLiion Robotic Cat development system. We’ll be offering kits (partially assembled, basic, complete, pre-built, etc.) via Indiegogo in the near future. In the mean time, we’ll be blogging about the project here at SpaceSpeak. Our current plans call for launching this system under the AYRRUS brand (All Your Robots R US).
We’ll add additional details on this page as the project progresses…
History (and photos) in reverse chronological order (most recent stuff first)
9/23/2018 –> Up to version 24d. Added object tracking ala Pixy 2 (see Charmed Labs, amazon to get yours). Also changed the leg design and circuit board stack.
version 23 object tracking
9/30/2017 –> Numerous updates have been scattered throughout this site and on the SpaceSpeak FaceBook page. I’ve tried to pull them altogether here.
I’ve moved back to 180 degree servos. Moved to a different stronger version of the mg90 –> mg92b and have radically simplified the design. Here is the before and after simplification photos. Biggest functional change however was adding a 3rd servo to each leg to control leg splay. This allows better walking and the ability to add some shoulder roll dance moves.
Also reworked the tail a bit to improve tip twitch and integrate both servos into the tail itself. This improves performance and further simplifies the design. In addition, it allows me to use the tail design to create a cat tail hat – flashing ears and twitching tail running on a Raspberry Pi zero.
After Simplification and shoulder shrug servo addition:
Version 22 – new servos, shoulder servos, tail design, and battery board
Version 20 – simplified/ new 180 degree servos – no shoulder servos
1/30/2017 –> Switched to continuous rotation servos with potentiometers measuring actual rotation at each joint. Also added front and back leg rotations via a waist servo.
12/27/2016 –> Here are some videos showing functionality of the API. These are using a development test frame for the legs which allows me to flip the critter onto its back which isn’t possible with the actual robot. Also there aren’t any paws or tail installed, so it isn’t particularly stable in the upright position. That said, the test looks pretty good. Plenty of power in the servos, and coordinated movements are working just great.
12/23/2016 –> Software API is coming along nicely. I have threaded the individual servos (base level API) and have coordinated movement of the servos for each leg (normal walking and high step walking). Next is coordinating the movements of all four legs at once.
There are a number of gaits that a cat uses:
- The diagram above shows a normal walk. On the right side, both legs are touching the ground but aren’t straight up and down. This makes that side of the cat tilt down a bit as compared with the left side of the cat where the front leg is fully extended and vertical. This lateral tilt of the cat moves the center of weight toward the right hand side. This in turn allow the left rear foot to be lifted off the ground without tipping the cat over. The arrows indicate the direction each leg will move in the next moment. This gait typically has only one leg off the ground at any given time.
- Trot – when you feed your cat, this is the gait that he or she uses to get to wherever you will be depositing their food bowl. The tail is normally up for this mode of movement and the front legs run like a human. The back legs also pair up and run like a human. The two sets are out of phase so that diagonal legs are either up together or down together. This gait always has two legs off the ground.
- Full out gallop is the easiest to understand – Both front legs reach forward and pull back at the same time. The back legs work together as well – They push back and then while the cat is airborne, they reach forward to create as long a push zone as possible once they land again. This mode is used to accelerate quickly.
- Half gallop is a little less energy intensive. The back legs continue to work together; but the front legs start alternating. They don’t really do much pulling in this mode. Their real purpose is to smooth out the run and keep the keep the cat from doing a face plant in the dirt.
12/10/2016 –> Still working on software API. Diverted a bit this week to work on upgrading one of my 3D printers to print with four filaments in prep for the soon-to-arrive Polysmooth alcohol based smoothing machine and associated new line of high strength filaments. Also reached out to the team at Reach Robotics to see if we can partner with them. If you haven’t seen their system yet, check it out here. ’tis the season for cool stuff!
11/30/2016 –> Working on software now. I have the camera (still shot and video) working, as well as the tail wagging function.
11/19/2016 – version 11 –> added pin and slot steering for the front legs. Also changed raspberry pi mounting, upgraded tail gearing, and added motorized pan and tilt to front sensor platform (aka head/camera).
11/01/2016 – printing new body frame with 5% leg splay added for increased walking stability
we’ve reworked the tail so that it drapes more naturally. Looks more like a cat now instead of a lemur. Time to change modes and start working on control software again. Note the tail is more than decoration. We’ll need it as a dynamic counter balance during walking.
10/15/2016 – we’ve added a fully articulated tail (21 bones, 3 servos) with amazingly life-like movement. Also upgraded our leg servos (6x power increase), added new software, and a 3D camera system using a GoPro and a new Vitrima 3D lens.
We intend to start the www.Indiegogo.com campaign in the first week of November.
9/20/2016 – new feet. This is the last version with the MG90 servos. Moving to MG995 and MG996
9/01/2016 – Our summer Intern has returned to school. She made excellent progress on our first bio-inspired robotic cat. The cat physiology research, core design and implementation is her work.Here are two of the early alpha models based upon the Raspberry Pi 3.
And here are some random photos from the last three months