Ranger 2008           (Looking for Ranger 2011?)

This version of Ranger is a re-worked version of the robot which walked 1 km in December 2006. Changes include the addition of an IMU (inertial measurement unit), 3 satellite computers, 3 ankle-angle encoders, a hip angle encoder, a hip-angle spring, and more batteries. Additionally, the cable drive was balanced to have equal stiffness on inner and outer legs and the mass redistributed so that the inner and outer legs would be dynamically nearly equal. The ankle gear ratios were changed. The software was seriously reworked to incorporate various control principles which we think are important (as will be explained in a forthcoming paper).
Pre-emptive push off is still not possible. Cornell Ranger's specific energetic cost of transport has been reduced from 1.6 (December 2006) to 0.60 (April 2008). This is still much higher than the Collins robot (0.2), but still relatively low in the robot world. That is, the latest 54 kg Honda Asimo Humanoid uses about 1200 watts and can run at 6 km/hr. If it uses 1200 watts at full speed (can it literally run for an hour?) then it has a specific cost of tranport of about 1.3. Boston Dynamics BigDog is said to have a 25 hp motor and is about 250 kg fully loaded going perhaps 1.8 m/s. That gives a specific cost of transport of about 4 (or about 16, given 25% efficiency in converting gasoline to mechanical power). Humans have a specific cost of transport of about 0.2 (food energy per distance per weight). [A student was accidentally quoted in some press reports as saying that the Cornell Ranger uses 1000 times less energy than the Honda Humanoid. In fact the Cornell Ranger uses about 50 times less energy and is much lighter (and less versatile)]. The Cornell Ranger , while "passive" inspired, is not passive. The information flow rate is perhaps 10,000 times greater than for the Collins robot. The basic control loop runs once every millisecond and many quantities are measured with at least 8 bit accuracy. And all three joints are controlled. But the control philosophy is still pretty dumb: push off when sensing heelstrike, pick up the foot, swing the leg forward, put the foot down, wait for heel strike, repeat. We have found controller designs that use as little as 15 watts, however these are not sufficiently stable for long walks. If we can stabilize that walk, and use the full 160 watt-hours on our batteries Ranger could walk 18 km (over 11 mi). We'll see!! See many more pictures and videos and design sketches on the 2006 Ranger page

From a theoretical/scientific point of view this robot is less interesting than the Collins machine. Ranger has 4 legs instead of just two, no knees, a less natural gait and it uses more energy. But there is a key difference. Ranger works most days. Thus we can tune it and learn from it. The goal is to get back to the grace and low-energy use of the Collins robot, but with reliability. Meanwhile the Ranger seems to be a world record holder for autonomous legged robot distance.