To control movements aided by functional electrical stimulation (FES) in paraplegic patients, stimulation of the paralyzed lower limbs might be adjusted in response to voluntary upper body effort. Recently, Donaldson and Yu proposed a theoretical approach, called "control by handle reactions of leg muscle stimulation" (CHRELMS), in which stimulation of the lower limbs depends on upper body effort, i.e., body posture and recorded hand reactions, and is aimed to minimize arm forces during standing up and standing. An alternative strategy is presented in this paper, which accounts for voluntary upper body effort as well, but does not require estimation of hand reactions. The objective of this study is to test both strategies by applying them to a generic two-dimensional (2-D) neuromusculoskeletal model. The model takes into account the major properties of muscle and segmental dynamics during FES-supported standingup movements of a paraplegic patient. In comparison to standing up without FES-support, both closed-loop strategies yield satisfying standing-up movements although no reference information (e.g., a desired trajectory) is required. Arm forces can be significantly reduced. Using the model to optimize the controller, timeconsuming and strenuous trial-and-error experimentation could be avoided. However, final experimental studies are planned to verify the presented strategies.

The goal of this paper is to design WALK! a cooperative, patient-driven neuroprosthetic (NP) system. In implementing sensor-supervised events to switch to subsequent medical prosthetics, NP users were able to actively control the timing of their movements. Performance and usability of WALK! was appreciated by the NP users because they were able to perceive the activities of the NP to actually support their movements. The future of NP will be based on fully implanted systems. To justify the high efforts, risks, and costs of an implantation to both NP users and health care providers, NPs have to offer true functionality that can only be achieved by a sophisticated and yet practicable control system. We believe that the WALK! control approach presented in this article can be considered a valuable contribution to the development of future neuroprosthetic systems for locomotion.

No abstract

The use of mathematical models has the potential to enhance the development of lower extremity neuroprostheses (NP) based on Functional Electrical Simulation (FES). The choice of model complexity is not trivial when building a model for FES control design. On the one hand, a comprehensive model might be useful to account for the many different biomechanical and neurophysiological effects that can be observed during FES-induced movements. On the other hand, too complex models are difficult to be utilized in and identified for NP applications. In this paper we discuss the disadvantages of too complex models, and propose potential simplifications on the basis of existing models that are commonly used to describe muscle activation, muscle contraction and body-segmental motion. The obtained model approach is simple enough to be identified, and sufficiently comprehensive to describe most of the relevant effects that occur during FES-induced locomotion.

No abstract