DARPA Pushes Machine Learning with Legged LittleDog Robot

DARPA's LittleDog project is an effort to build an autonomous legged robot that is aware of its environment and capable of deliberately placing its feet to avoid falling. The software used in LittleDog determines the robot's route and its cameras and leg sensors help it detect obstacles to avoid missteps. DAPRA wants LittleDog, a follow-up to its BigDog project, to be able to move across progressively more difficult terrain at increased speeds. "BigDog and LittleDog are related in that they are both focused on solving the problems that will enable legged robots to accompany war fighters as they cross complex terrain," says DARPA's Tom Wagner. Phase three of LittleDog's development process is scheduled to begin this summer. Phase one challenged six teams of roboticists to improve on the basic robot platform developed for BigDog. Successful completion of phase one required each team's LittleDog to move at a rate of at least a half inch per second over terrain that included obstacles 1.9 inches high. Read More