Figure 1.

Adaptive embodied neural closed-loop control setup and the real and simulated biomechanical walking robots AMOSII. (A) The setup of adaptive embodied neural closed-loop control for complex autonomous behaviors with versatility and adaptivity. It consists of the adaptive neural sensory processing network with synaptic plasticity, the modular neural locomotion control network, the biomechanical walking robot AMOSII, and the environment. The sensory processing network is in the focus of this study while the locomotion control network and the robot have been developed earlier (Manoonpong et al., 2013). (B) The real biomechanical walking robot AMOSII with 19 DOFs and its sensors. It has one backbone joint for climbing and 3-jointed legs for walking on different terrains. Its two ultrasonic sensors installed at the front are used to provide exteroceptive sensory feedback to its adaptive neural sensory processing network. Additionally, a camera and a laser range finder are used for terrain recognition and obstacle height detection. (C) The simulated AMOSII using the LPZRobots simulation environment (Martius et al., 2010).