Research: Bionics

Neuro-Muscular Electrical Stimulation

Neuromuscular Electrical Stimulation (NMES) is a technique used to generate desired muscle contractions via electrical stimulus and has potential to facilitate improved limb control and functionality for patients with spinal cord injury, stroke, and neurological disorder. The purpose of NMES research in the nonlinear control and robotics laboratory is to develop nonlinear control methods for artificial stimulation of human skeletal muscle. Our research focuses on using different computer-controlled stimulation methods to enable a human knee joint to follow a desired trajectory. Electrodes are placed on a human quadriceps muscle and an electrical signal is applied to either induce a contraction to extend the leg to a specified degree or have the leg follow a specified trajectory. This research is motivated by the desire to augement human performance through machine interaction. The image depicts the NCR Masters student Keith Stegath undergoing Neuro-Muscular Electrical Stimulation. Collaborative multidisciplinary efforts are being coordinated with the UF Department of Physical Therapy and the VA-Brooks Human Performance Laboratory, Gainesville, FL.




Biomimetic Approaches to Flight Control and Inertial Stabilization

Natural systems have evolved very robust solutions to highly complex problems. Large numbers of simple low fidelity sensors are typical, rather than sophisticated highly optimized single sensors. Sensor fusion, deriving measurements from multiple sensor modalities is common, for example, using both visual and inertial sensing simultaneously to measure body angular rate. The goal of this research is to learn from nature robust solutions to feedback control problems that might not have normally been considered following the traditional paradigm of engineering design. Current activities are focusing on insects to understand how inertial measurement occurs for flight stabilization. Gyroscopic sense organs called halteres are being simulated to understand how insects reconstruct 3 orthogonal components of the inertial rate vector using very simple fields of strain sensors. 6DOF flight simulations are also being constructed to reveal how extremely dynamic flight maneuvers and complex navigation solutions result from fusion of visual, inertial, and olfactory feedback.




Evolutionary Algorithms for Heuristic Adaptation in Non-Linear Control

Cumulative selection is the process through which natural systems evolve over long periods of time within changing environments. Through this process natural systems adapt to long term changes in weather, develop resistance to disease and predation, and adapt to changes in food supply. Engineers commonly use cumulative selection in the form of genetic algorithms to train neural networks or to find optimal solutions to problems for which exact solution techniques are intractable (e.g., NP-hard problems). This research area intends to develop new techniques that continually evolve optimal non-linear control parameters in environments where the plant and the control criteria are changing with time.