Quaternion Spiking Neural Networks Control for Robotics

Abstract

In this work is implemented an adaptive controller for robotics systems, where the adaptive control is using spiking neural networks in the quaternion algebra framework. The main advantage of our controller is the mathematical modern language we use and that the algorithm can be applied to any robot architecture. In other words, the designed controller is developed for robots with any degree of freedom (DoF).Another advantage is the opportunity to solve the inverse kinematics for any robot using screw theory, this advantage is important for redundant robots. Where traditionally, in redundant robots, the solution of the inverse kinematics can be complicated or tedious to solve.