This research involves recognizing prosthesis user activity and intent (for example, standing, walking slow, walking fast, going up stairs, going down stairs, walking up a ramp, walking down a ramp, and so on). Prosthesis control parameters are dependent on user activity, so it is important to recognize user activity, and to quickly switch control modes when the user intends to change his gait mode. User activity and intent recognition involves optimal filtering, clustering, and artificial intelligence.

​

We designed an accurate and parsimonious user intent recognition (UIR) to control seamless transition between low-level controllers depending on the user’s intention and environment. We developed a new hybrid evolutionary algorithm, known as invasive weed optimization/biogeography-based optimization (IWO/BBO), for optimizing UIR performance and design parameters. Our hybrid algorithm featured three new components that are typically not present in standard IWO: (1) migration; (2) gradient descent; and (3) mutation. These novel operators strengthen the optimizer exploration and exploitation capability. We used non-invasive signals from the prosthesis and the residual limb as the relevant input signals for UIR. Experimental data for training and testing our system were collected at the Cleveland Department of Veteran Affairs Medical Center from able-bodied and transfemoral amputee subjects. An advanced state-of-the-art facility with a V-Gait treadmill that uses motion sensors and cameras were in place to capture body movement. Multiple walking trials were collected from each subject, where each trial was a sequence of various gait modes. Various time-domain features were extracted. Principal component analysis was used to eliminate the least relevant features. We trained a multi-layer perceptron artificial neural network and a modified K-nearest neighbor classifiers using hybrid IWO/BBO. Then the optimized UIR system was implemented to identify unknown walking activities in real-time, and achieved a high accuracy of 96%. Wilcoxon statistical tests showed competitive performance of hybrid IWO/BBO with comparison to the 10 other state-of-the-art evolutionary optimization algorithms.