About AmoebaCopters

For now the only person involved directly is me. However, since this initiative is very new, more people are likely to join. My idea was inspired while working in an MSc group project at The University of Sheffield, where we had to build a rotorcraft using not more than three motors. I tried developing an LQR-based controller for a weird-looking tricopter (uploaded as TricXControlArch, and is open-source), which my team was supposed to use. I identified the problems eventually after a test with JSBsim, like yaw control, and tried developing an LQR-PD combination and eventually a full 12-state LQR, both being failures. However, after I developed an INDI system for customized tricopters, it seemed to work, atleast on simulation. These, however, were not a direct part of the group project, and was completely developed by me in an independent manner, and has not been physically tested with a real copter, so I would strongly recommend against using any of this in a real-life system unless I thoroughly test everything successfully in the real world (which would for now include building 3D printed drones and integrating my autopilot system, flying first tethered and then freely). My eventual goal is to create an aerospace ecosystem for dynamic scaling, using copters as testbeds. While ugly copters won't fly well, they can be suitable testbeds for control systems. I am currently building a physical laboratory for AmoebaCopters.

The entire mathematics is done by me, while software integration like C++ coding and testing is done by Claude. I also use Claude to analyse test results, of course with my own understanding. This is well documented in the commit history. I take full responsibility, but I would thank Claude by Anthropic for coding and management of files and folders, and making computers easy for everyone.