Working towards a highly mobile, legged exploration robot, the DLR Crawler is a first experimental platform to test various control and gait algorithms as well as different approaches to vision-based navigation. The robot employs the fingers of DLR Hand II as legs, which offer a unique combination of high performance actuation and a broad range of integrated sensors.
The robot was first presented to the public in 2008.
technical data
Size:
footprint: 35 cm x 38 cm, ground clearance: up to 12 cm
Mass:
3.66 kg
Degrees of freedom:
18 active leg joints
Payload:
3 kg
Speed:
up to 20 cm/s
Power supply:
external 24 V supply
Computation:
external control and navigation PCs; fast SpaceWire link closes 1kHz control loop
Sensors:
• joints: motor-side Hall effect sensor, link-side potentiometer and torque sensor, temperature sensor • foot: 6-DOF force-torque sensor • body: IMU, stereo camera
User interface:
• spacemouse for direct control • autonomous waypoint navigation to predefined goal
System description
Aiming at terrestrial and extra-terrestrial exploration, all locomotion- and navigation-related algorithms developed for the DLR Crawler focus on robust autonomous operation in a priori unknown, GPS-denied environments. Based on findings from biology, the gait algorithm of the DLR Crawler employs a highly flexible, decentralized approach, wherein neighboring legs influence each other by inhibiting or exciting stepping motions. Thus, the gait of the DLR Crawler does not show a fixed pattern but rather emerges according to the commanded velocity and the interaction of the legs with the local environment. In addition, joint torque sensors provide valuable proprioceptive measurements that enable sensitive interaction with the terrain by compliance control. Thus, in combination with three simple reflexes, the robot is able to negotiate all obstacles within its walking height autonomously. Furthermore, the gait coordination immediately adapts to the loss of a leg while the robot easily maintains its stability. In order to navigate in unknown terrain a stereo camera head is mounted on the DLR Crawler. Using the semi-global-matching-based stereo vision algorithm developed at the institute, the robot assesses the terrain traversability and plans a path to a desired goal solely based on the information acquired by its on-board sensors.
M. Görner et al., “A leg proprioception based 6 DOF odometry for statically stable walking robots” in Autonomous Robots, 34 (4), pp. 311–326, Springer, 2013.
A. Stelzer et al., “Stereo-Vision-Based Navigation of a Six-Legged Walking Robot in Unknown Rough Terrain”, in the International Journal of Robotics Research, Special Issue on Robot Vision, 31(4), pp. 381–402, 2012.
A. Chilian et al., “Multisensor Data Fusion for Robust Pose Estimation of a Six-Legged Walking Robot”, in IEEE/RSJ International Conference on Intelligent Robots and Systems, San Francisco, CA, USA, 2011.
M. Görner et al., “Analysis and Evaluation of the Stability of a Biologically Inspired, Leg Loss Tolerant Gait for Six- and Eight-Legged Walking Robots”, in IEEE International Conference on Robotics and Automation, pp. 1525–1531, 2010.
M. Görner et al., “The DLR Crawler: evaluation of gaits and control of an actively compliant six-legged walking robot”, Industrial Robot: An International Journal, 36(4), pp. 344–351, 2009.
M. Görner et al., “The DLR-Crawler: A Testbed for Actively Compliant Hexapod Walking Based on the Fingers of DLR-Hand II”, in IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 1525–1531, 2008.
