Biomimetic Control of Traditional Prostheses

Human joints are not stiff, even when people tense their joints. This low level of stiffness allows people to move quickly without fear of breaking limbs or objects in their environment. In contrast, electrical prosthesis joints are very stiff. This stiffness can cause gears in the prosthesis to break if the user falls. As more powerful motors are introduced in more proximal prosthetic joints, the authors have found that the speed of the prosthesis has to be limited to ensure safety to the user due to the stiffness of the prosthesis. Ideally, electrical prostheses would be less stiff in order to increase their performance while ensuring safety.

Low levels of stiffness have been advocated since Hogan suggested a control paradigm termed impedance control. In impedance control, low levels of stiffness are varied by the user depending on the task (1). For many reasons, it is difficult in the realm of prosthetics to create a “soft” prosthetic joint, and as a result, a practical implementation of Hogan’s control paradigm was not found until the mid 1990’s. This solution, termed a series elastic actuator by Pratt et al. (2), considers the physical characteristics and the microprocessor characteristics of the motor at the same time. By shifting the lack of stiffness from the microprocessor to the actual physical device, promising results have been achieved. Introducing a spring at the output of the motor prevents large forces from being generated when the prosthesis collides with an object. It negatively affects performance by decreasing the maximum speed at which the prosthesis can generate large forces.

Our research has attempted to harmonize the desired features of series elastic actuators, namely their inherent safety and accuracy, with a conservative power source required in prosthetics. This has traditionally been done by the inclusion of a non-backdrivable transmission, which only draws energy when movement of the prosthesis occurs. Nonbackdrivable transmissions have not been used in series elastic actuators, but there are no inherent conflicts between their use and safe, accurate control. We have found that the responsiveness of the prosthesis is not reduced by the inclusion of a non-backdrivable transmission. We are currently integrating these concepts into a prosthetic elbow shown in figure 1. A more detailed discussion may be found in (3).

This non-backdrivable series elastic actuator will allow users to operate the prosthesis using Hogan’s biomimetic control paradigm, in which the stiffness of the prosthesis and the position of the terminal device are controlled. As a result, the prosthesis will automatically compensate for unexpected collisions, just like an able bodied joint. By wrapping the spring through the middle of the motor and using a non-backdrivable gear transmission, this design improves the safety and performance of electrical prosthetic joints while preserving the modularity, size constraints, and power conservation required in prostheses.

This work was supported in part by the National Defense Science and Engineering Graduate Fellowship; the Department of Veterans Affairs, Rehabilitation Research and Development Service administered through the Jesse Brown VA Medical Center, Chicago; and the National Institute of Disability and Rehabilitation Research of the United States Department of Education under grant H133E980023. The opinions in this paper are those of the authors and do not necessarily reflect those of the Department of Education.

Figure 1: Nonbackdrivable Series Elastic Actuator in an elbow

1. N. J. Hogan, "Prostheses should have adaptively controllable impedance," presented at IFAC Control Aspect of Prosthetics and Orthotics, Ohio, 1982.

2. G. A. Pratt and M. M. Williamson, "Series elastic actuators," presented at IEEE/RSJ International Conference on Intelligent Robots and Systems, Pittsburgh, PA, 1995.

3. J. W. Sensinger, "Design & Analysis of a Non-backdrivable Series Elastic Actuator for use in prostheses," in MSC Thesis, Biomedical Engineering. Evanston: Northwestern University, 2005, pp. 135.