Rhex - one of the most successful designs of small six-legged robots. His legs have no joints and are arcuate springs that rotate around their axis, so only six actuators provide the mobility of the robot. Rhex can run at speeds of up to 2.7 meters per second with a body length of about half a meter, travels almost 4 kilometers over a harsh terrain on a single battery charge, climbs a hill with a slope of 45 degrees, swims and climbs the steps.
One of the latest achievements of Rhex is jumping. Jumping robot taught in the laboratories of the University of Pennsylvania. He jumps over the “abyss” with a width of 50 cm from the spot and 60 cm from the takeoff run (or rather, from the jump), makes a backward flip, jumps onto an obstacle 29 cm high, and clings to the ledge up to 73 cm in height, and then climbs onto it from hanging position, though it is given to him with great difficulty - his legs are not designed for climbing.
The Rhex project was launched in 1998 on the initiative of DARPA, about $ 8 million has already been invested in the robot. Rhex is engaged in the development and improvement of several large universities and the company Boston Dynamics - the creator of walking, jumping and running military robots Big Dog, Petman, Cheetah and others. ')
Unlike Boston Dynamics, which does not publish virtually any details about the achievements of their robots, researchers from the University of Pennsylvania published a detailed article on how they managed to achieve such results. Scientists have analyzed the space of possible states of the robot, from complete rest, when all its legs rest on the ground, to states in which the robot does not touch it at all, and have compiled a “dictionary” of possible robot movements that lead to a jump, that is, a transition map from one state to another. Based on this vocabulary, they programmed several jumps. In the future, this approach will allow us to formally explore the entire space of the robot's movements and discover effective movement strategies that have no analogues in nature and are not limited by the imagination of scientists. This will help make the robot faster, more efficient and more versatile.