Evolving stability in bipedal robotic systems

In order for a bipedal robotic system to interact with its environment it must be able to have stability. And because an environment can be composed of different types of terrain and surfaces, adaptation must be part of system behavior. Many current bipedal systems (e.g. Honda´s P3 and Asimo or Sony´s) utilize high torque custom designed servos and/or control systems to handle specific terrain (flat surfaces). What remains to be seen is how well these bipedal systems respond to terrain that isn´t well defined. A robust stable system will be free of behavior that could cause damage to the system such as can occur in a fall. This thesis proposal proposes this. In order to have a completely stable system, a biped robot must be able to adapt to its environments terrain in real time and have the ability to employ corrective behavior in the event of an occurrence of instability.