Hexapod Gait Adaptation Based on CPG in Case of Limb Damage

The integration of ideas from spiking neural networks and reinforcement learning (RL) algorithms is a promising direction in mobile robotics, as it allows for more energy-efficient solutions capable of handling larger state and action spaces. One of the important tasks is adapting to changes in the environment and the state of the mobile robot. This work proposes an adaptation algorithm for a hexapod's gait to limb damage, which is based on CPG (Central Pattern Generators) and reinforcement learning. This enables the organization of high-level logic through simple computations, which could be beneficial for the use of more complex algorithms in the future.