The purpose of this paper is to present and evaluate methods of control and gait generation for the DLR Crawler – a six‐legged walking robot prototype based on the fingers of the DLR Hand II.
Following the institutes philosophy, the DLR Crawler is a highly integrated mechatronic device. As in all DLR robots, joint torque sensing plays an important role to allow actively compliant interaction with the environment. To control the Crawler a joint compliance controller is implemented and two different methods of gait generation are in use. The first method, intended for moderately uneven terrain, employs scalable patterns of fixed coordination combined with a leg extension reflex. For the second method, used in rougher terrain, a set of rules found by biologists in stick insect studies is applied. Based on these rules gaits emerge according to a velocity command. These gaits are combined with several reflexes to a reactive walking algorithm.
The compliance controller together with the reactive gaits allows the robot to autonomously master uneven terrain and obstacles with height differences within the nominal walking height. Further, the controller reduces internal forces compared to pure joint position control. The sensitive joint torque sensors allow fast collision detection and reactions thereafter.
This paper introduces a six‐legged walking robot test bed with comprehensive force‐torque sensing capability. Joint compliance controllers are implemented and successfully combined with reactive gait algorithms. For the second gait algorithm inspired by Cruse's rules, which were identified for forward walking stick insects, an implementation has been found for the DLR Crawler that gives the robot full omnidirectional mobility.
