The insect brain is a very efficient neural computing system.
It can process high-dimensional sensory information and generate
coordinated and adaptive motor commands in real time,
resulting in various complex behaviors (including locomotion, object
manipulation, navigation, and their combination). Simultaneously,
it can also autonomously learn to solve complex tasks. This amazing
control performance is achieved by using the full capacity of its neural
dynamics, learning, memory, and plasticity as well as by interacting
with the environment through its body (i.e., embodiment).
Inspired by this, we have developed brain-like mechanisms. The mechanisms are based on a modular concept and hierarchically organized. They exploit neural dynamics, learning, memory, and plasticity, as the biological brain does, to efficiently generate complex functions of embodied multi-sensorimotor robotic systems. Based on
this development, we have addressed the way to achieve adaptive motor control and autonomous learning principles for complex locomotion, object manipulation, goal-directed navigation, and their combination in the embodied systems.