The goal here is to develop a powered bipedal robot that walks in a more human-looking manner and with smaller energy cost than all previous walking robots that have similar gross abilities. The geometry of the robot will be loosely based on human body parts. This robot will generate its life-like motions by chasing efficiency and simplicity rather than by tracking trajectories. Others have built beautifully engineered humanoid walking robots with flat feet and with motions that are carefully controlled to move smoothly, joint by joint and at every instant in time. These robots typically use about 50 times as much energy per unit weight and distance as does a person. Ruina's approach will build on his Cornell lab's recent bipedal robot that has a natural-looking gait and which uses (scaled) about as much energy as is used by a human. The existent Cornell robot, based on the `passive dynamic' approach of McGeer, is basically sticks and hinges arranged in a way that is reminiscent of a human body, and with a small motor to extend the ankles as each foot leaves the ground. The robot being developed will add functionality including the ability to stand still, to start and stop walking, to walk at various speeds, and to turn. Intellectual merit: The proposed new robot will test the general hypothesis that robotic human mimicry is well-accomplished by holding energetic efficiency and control simplicity paramount. Broader impact: Robots serve as a natural vehicle of science education. As in past years at Cornell, undergraduate students will be involved in the research. Visiting high-school students are inspired by visits to the lab. Cooperation with the Science Museum of Minnesota Cyborgs exhibit will continue as will cooperation with the media. Dissemination of the should alter the mode of thinking of those involved in the design of robots and also of those involved in the diagnosis and correction of human walking disabilities