Objective - The objective of this study is to implement and test a neuro-mechanical gait assist (NMGA) device to correct gait characterized by muscle weakness, incoordination or excessive tone in veterans with hemiplegia from stroke that adversely affect their ability to walk, exercise, perform activities of daily living, and participate fully in personal, professional and social roles. Research Plan – Prototype NMGA device will be used to develop a finite state controller (FSC) to coordinate user’s volitional effort coordinated with surface muscle stimulation and motorized knee assist as needed. Brace mounted sensors will be used to develop gait event detector (GED) which will serve the FSC to advance through the phases of gait or stair climbing. In addition, a supervisory rule-base intent detection algorithm will be developed using brace mounted sensors and user interface input to select among various functions including walking, stairs climbing, sit-to-stand and stand-to-sit maneuvers. The FSC controller tuning and intent algorithm development and evaluation will be on three pilot subjects with gait deficits from stroke. Outcome measures during development will provide specifications for a new prototype NMGA design which will be evaluated on three pilot subjects to test hypothesis that NMGA improves walking speed, distance and energy consumption of walking. These baseline data and device will be used to design a follow-up clinical trial to measure orthotic impact of NMGA on mobility in activities of daily living at home and community. Methodology – After meeting inclusion criteria, pilot subjects will undergo baseline gait evaluation with EMG activities of knee flexors and extensors, ankle plantar and dorsiflexors and isokinetic knee strength and passive resistance. They will be fitted with NMGA combining knee-ankle-foot-orthosis with motorized knee joint and four channels of surface neuromuscular stimulation of plantar- and dorsi- flexors, vasti and rectus femoris. Brace mounted sensor data will be used for gait event detector (GED) algorithm development and evaluation. GED will serve FSC to proceed through phases of gait based on supervisory rule-based user intent recognition algorithm detected by brace mounted sensors and user input interface. FSC will coordinate feed-forward control of tuned stimulation patterns and closed-loop controlled knee power assist as needed to control foot clearance during swing and stability of the knee during stance. Based on data attained during controller development and evaluation, a new prototype NMGA will be design, constructed and evaluated on three pilot subjects to test hypothesis that NMGA device improves safety and stability, increases walking speed and distance and minimizes user effort. For all experimental conditions, a statistical power of 90% was used to determine the number of repetitions required in the single subject design to detect statistically significant or 0.2m/s minimal clinically import...