Description
This research thread is pursuing the ambition to bring the Bio-inspired Behavior-Based Bipedal Locomotion Control (B4LC) to a physical platform. Based on the experiences gathered by the deployment of B4LC to the simulated biped, the requirements on a physical system could be well defined. In order to render natural-looking and energy-efficient bipedal walking, a main concept of B4LC is to exploit the passive dynamics of the underlying system. Thus, the system should as far as possible imitate its anthropomorphic counterpart - regarding e.g. the actuation, the kinematic layout, and the weight distribution.
As a first iteration, the Compliant Robotic Leg (CARL) has been developed. It is a planar robotic leg that features mono- as well as bi-articular actuation. As the actuation is a key component in a robotic leg, a series of linear Series Elastic Actuators - the RRLab SEAs - has been developed. The design was mainly driven by two requirements: the capability to act as a force/impedance source and a inherent tolerance against impact forces. Each RRLab SEA is encapsulated by a dedicated FPGA-based system. For more information see the dedicated page linked above.
Within the leg the SEAs are acting on the joints either by direct or by four-bar linkages. This resulted in redundant system in which all five SEAs are coupled. Motivated by the scientific evidence that human amputees are well capable of producing natural walking using a combination of a SEA and a off-the-shelve prosthetic foot, an ots product is used for the leg as well.
To enable a walking motion with a single leg a test rig incorporating a treadmill and a lifting mechanism has been developed. Especially the latter is important as it allows for the imitation of a second, virtual leg.
After validating the low-level impedance and force control even the case of the fully coupled system, a first walking motion could be generated. Therefore, a subsystem of B4LC has been ported to the leg.
Publications
- Continuous Inverse Kinematics in Singular Position.
Robotics for Sustainable Future, S. 24 - 36. (2022) - Biologically Inspired Bipedal Locomotion - From Control Concept to Human-Like Biped.
Proceedings of 14th International Conference on Electromechanics and Robotics “Zavalishin’s Readings”, S. 3 - 14. (2020) - CARL–A Compliant Robotic Leg Designed for Human-Like Bipedal Locomotion.
(2020)
https://kluedo.ub.uni-kl.de/frontdoor/index/index/docId/5975 - Exploiting the intrinsic deformation of a prosthetic foot to estimate the center of pressure and ground reaction force.
Bioinspiration & Biomimetics, (2020) - Integration and Design of Actuation Redundancy in Robotic Leg CARL Based on the Physiology of Biarticular Muscles.
Dissertations – Technical University of Kaiserslautern, (2020)
https://www.dr.hut-verlag.de/9783843945981.html - Integration and Design of Actuation Redundancy in Robotic Leg CARL Based on the Physiology of Biarticular Muscles.
(2020) - SLIP-Based Concept of Combined Limb and Body Control of Force-Driven Robots.
Advances in Service and Industrial Robotics, Vol. 84, S. 547 - 556. (2020) - Design of the musculoskeletal leg based on the physiology of mono-articular and biarticular muscles in the human leg.
Bioinspiration & biomimetics, Vol. 14, Nr. 6, S. 066002. (2019) - FPGA-based Embedded System Designed for the Deployment in the Compliant Robotic Leg CARL.
Proceedings of the 16th International Conference on Informatics in Control, Automation and Robotics - Volume 2, S. 537 - 543. (2019) - Coordination of the Biarticular Actuators Based on Instant Power in an Explosive Jump Experiment.
IEEE International Conference on Advanced Intelligent Mechatronics (AIM), (2018) - Moment Arm Analysis of the Biarticular Actuators in Compliant Robotic Leg CARL.
Conference on Biomimetic and Biohybrid Systems, S. 348 - 360. (2018) - CARL – A Compliant Robotic Leg Featuring Mono- and Biarticular Actuation.
IEEE-RAS International Conference on Humanoid Robots, S. 289 - 296. (2017) - Modular Control Architecture for Bipedal Walking on a Single Compliant Leg.
(2017) - Off-Road Path Detection Using Convolutional Neural Networks.
(2017)