Hydrostatic actuation for wearable robotics
Kyrian Staman is a PhD student in the Department Biomechatronics and Rehabilitation Technology. Promotor is prof.dr.ir. H. van der Kooij from the Faculty of Engineering Technology.
Exoskeletons require energy efficient actuators of low mass and volume (to increase wearability) with controlled output force, sufficient for full gait support. The main research objective of this work was to develop such an actuator.
First, the concept of the passive return electro-hydrostatic actuator (PREHydrA) was developed. It eliminated conventional hydraulic systems’ fluid supply and valves, potentially making it lighter, more efficient, and simpler. It also avoided the configuration-dependent friction of Bowden cable transmissions. A physical port-based network model was created of the PREHydrA that predicted force tracking with a maximum error of about 4 N. Closed loop output force control was used in tabletop experiments to obtain a mean absolute tracking error below 4 N for force references from 300 N amplitude at 0.5 Hz to 20 N amplitude at 10 Hz. These forces, frequencies and corresponding velocities (up to 0.47 m/s) demonstrated that the PREHydrA's performance was suitable for use in lower limb exoskeleton technology (and other wearable applications).
Next, a preliminary design for the PREHydrA was tested to requirements for gait restoration; one of the most demanding tasks for exoskeletons. While the concept offered good wearability properties, it had never been used in exoskeletons for full support. Custom and small commercial components were used in a design for the knee joint. An experimental setup with a pendulum representing a swinging lower leg was used to show force and angle tracking performance. The results of zero interaction force (-400 to 1100 N actuator force range) tracking experiments were maximum mean absolute errors of 61 N (6.79 Nm joint torque error) at 5.5 Hz excitation and a full swing (70 °) within 0.35 s (0.8 m/s actuator velocity). These matched or exceeded current state of the art exoskeleton actuation and control performance.
Finally, a small-scale electro-hydrostatic actuator (the miniHydrA) was developed based on gait requirements and integrated into an ankle exoskeleton. Leakage and friction were identified for the hydraulic cylinder. Furthermore, a control strategy was proposed and designed. The prototype exoskeleton was evaluated on joint torque tracking capabilities for torques between 0 and 120 Nm in benchtop tests and in use by subjects walking on a treadmill. Results showed zero torque tracking performance with a mean absolute error of 0.03 to 2.26 Nm over a range of frequencies from 0 to 5 Hz in benchtop tests. Torque tracking errors between 0.70 and 0.95 Nm were observed during treadmill walking experiments, where none of the three subjects showed deviations from their normal ankle joint kinematics. These results showed equal or better performance than other actuator concepts, like electromechanical or Bowden cable systems.
The main contribution of this work is the development and application of an existing, but in exoskeletons almost unused, actuator concept for wearable robotics. This work shows how to model and design the components of this concept to create actuator systems for the use case and specifications presented (full support of human walking). Where previously full support exoskeletons, using remote actuation, were designed predominantly with mechanical transmissions (rods, Bowden cables or shafts), now the additional option of the hydrostatic transmission can be considered, with this work being an example of how to design it, what performance can be expected and what limitations can be encountered.
