Continuous Classification of Locomotor Transitions Performed Under Altered Cutting Style, Complexity and Anticipation

摘要: Intent recognition strategies for lower-limb assistive technologies (e.g., prostheses and exoskeletons) are typically limited to transitions from one terrain another level ground stairs) occurring under anticipated cognitive states in a straight path. However, such systems might be unable robustly handle unanticipated those that do not occur line, posing high fall injury risks individuals suffering and/or physical impairments. Understanding how classification of human locomotion is influenced by changes direction user anticipation important. In this study, we examined the performance linear discriminant analysis continuous straight-line walking, 45°changes (crossover sidestep cuts) cuts stair-ascent (i.e., “mixed” transitions), performed varied anticipatory conditions able-bodied individuals. Accelerographic gyroscopic signals lower- upper-body were used as signal inputs. Relatively accuracy levels observed during cut-to-stair, compared cuts. addition, maneuvers was differing amounts training data. It more than two bouts given transition needed provide significant improvement mixed transitions. substantial (~5 bouts) information deliver improved (without stair). This study has implications enhancing design adapt cognition. These outcomes could inform robust handling “unknown” circumstances an increased safety confidence.