I would like to make a correction to this post. I had made an incorrect statement above, which appears in the following line of the article used above.
' "We use a good 1,000 times less energy than [Honda's] Asimo," Hippalgaonkar says. '
This is factually incorrect. This is incorrect, and was a mistake on my part. The Honda humanoid uses
about 10 times more energy per unit distance and weight, than the Cornell Ranger. It uses perhaps 50 times more power, in total. The error was solely mine and does not reflect anyone else's views in our lab ('The Biorobotics and Locomotion Lab' at the Department of Theoretical & Applied Mechanics, Cornell University).
* Our robot did finally manage to break its own unofficial distance-walking record. On April 3rd 2008, it exceeded even our expectations and walked 9 km (more than 5.5 miles) on a single charge of batteries, all the while using only 15-30 W of energy from its batteries all the time.
* Additionally here is some more information based on the Honda ASIMO website that might help compare the energy consumption and walking ability to theirs a little better. The most recent ASIMO has 6 kg of Li-ion batteries and operates 60 minutes on a full charge. Li-ion batteries have an energy density of up to 200
watt-hours/kg, and assuming that the Asimo has the best available, it would have about 1200 watt-hours available, and use 1200 watts. If the Honda robot could "run" for the full hour, and we don't know if this is
possible, it could go about 6 km in this time.
Source: http://asimo.honda.com/downloads/pdf/asimo-technical-faq.pdf