Serial servos: Lx-16a
Chess engine: Stockfish (https://stockfishchess.org/) with python API (https://pypi.org/project/stockfish/)
Computer vision: Potential dataset: https://github.com/maciejczyzewski/neural-chessboard https://par.nsf.gov/servlets/purl/10099572
Inverse Kinematics (kinematic coupling): https://www.researchgate.net/publication/261281825_A_screw_dual_quaternion_operator_for_serial_robot_kinematics https://www.researchgate.net/publication/276159627_Using_cuckoo_optimization_algorithm_and_imperialist_competitive_algorithm_to_solve_inverse_kinematics_problem_for_numerical_control_of_robotic_manipulators
All the good stuff related to kinematics (and robotics in general) is in this book:
Robotics: Modelling, Planning and Control by Bruno Siciliano, Lorenzo Sciavicco, Luigi Villani,Giuseppe Oriolo
Slowly getting to grips with 3D printer and putting it together. I've just managed to put together the first 3 joints and the servos for the first 2 degrees of freedom:
Created a 3d model of the robot and used what it seems to be the standard for their simulations. At a first glance it looks medieval compared to multibody simulations in other industries.
It's been a while since Kaggle allowed me to work on this project. This is a great source on QP, simple, clean and well explained (https://scaron.info/teaching/inverse-kinematics.html)
Implementation of this paper where decision variables of the controller are the incremental angle of the joints rather than speeds which seems to be the norm. Least square problem is transformed to QP problem with two inequality constraints, the incremental joint angles are limited by a max/min joint angles and max/min incremental values.
https://roam.me.columbia.edu/files/seasroamlab/publications/humanoids2013.pdf
