Connection between CPG model and servomotor of biomorphic fish robot

Most animal species locomotion is based on neuronal structures that called central pattern generators (CPG). Locomotor control carried out using such structures has some beneficial features that are useful to apply in robotics. However, the output signal generated by neurons is of a spiking nature, which is suitable for muscle control in animals, but not so suitable for controlling robot servomotors. In this work, we propose a method to connect the neural activity of a spiking neuron CPG model with the propulsion of a biomorphic fishlike robot. The signal transmitted to the robot’s servo is a sinusoid-like curve obtained by converting the muscle contraction variables of our model. Such a signal is well suited for servo control, and at the same time retains such features of CPG as resistance to short-term disturbances in network dynamics and the ability to smoothly modulating.