Versatility is important for a wearable exoskeleton controller to be responsive to both the user and the environment. These characteristics are especially important for subjects with spinal cord injury (SCI), where active recruitment of their own neuromuscular system could promote motor recovery. Here we demonstrate the capability of a novel, biologically-inspired neuromuscular controller (NMC) which uses dynamical models of lower limb muscles to assist the gait of SCI subjects. Advantages of this controller include robustness, modularity, and adaptability. The controller requires very few inputs (i.e., joint angles, stance, and swing detection), can be decomposed into relevant control modules (e.g., only knee or hip control), and can generate walking at different speeds and terrains in simulation. We performed a preliminary evaluation of this controller on a lower-limb knee and hip robotic gait trainer with seven subjects (N = 7, four with complete paraplegia, two incomplete, one healthy) to determine if the NMC could enable normal-like walking. During the experiment, SCI subjects walked with body weight support on a treadmill and could use the handrails. With controller assistance, subjects were able to walk at fast walking speeds for ambulatory SCI subjects—from 0.6 to 1.4 m/s. Measured joint angles and NMC-provided joint torques agreed reasonably well with kinematics and biological joint torques of a healthy subject in shod walking. Some differences were found between the torques, such as the lack of knee flexion near mid-stance, but joint angle trajectories did not seem greatly affected. The NMC also adjusted its torque output to provide more joint work at faster speeds and thus greater joint angles and step length. We also found that the optimal speed-step length curve observed in healthy humans emerged for most of the subjects, albeit with relatively longer step length at faster speeds. Therefore, with very few sensors and no predefined settings for multiple walking speeds or adjustments for subjects of differing anthropometry and walking ability, NMC enabled SCI subjects to walk at several speeds, including near healthy speeds, in a healthy-like manner. These preliminary results are promising for future implementation of neuromuscular controllers on wearable prototypes for real-world walking conditions.

Active prosthetic and orthotic devices have the potential to increase quality of life for individuals with impaired mobility. However, more research into human-like control methods is needed to create seamless interaction between device and user. In forward simulations the reflex-based neuromuscular model (RNM) by Song and Geyer shows promising similarities with real human gait in unperturbed conditions.The goal of this work was to validate and, if needed, extend the RNM to reproduce human kinematics and kinetics during walking in unperturbed and perturbed conditions. The RNM was optimized to reproduce joint torque, calculated with inverse dynamics, from kinematic and force data of unperturbed and perturbed treadmill walking of able-bodied human subjects.Torques generated by the RNM matched closely with torques found from inverse dynamics analysis on human data for unperturbed walking. However, for perturbed walking the modulation of the ankle torque in the RNM was opposite to the modulation observed in humans. Therefore, the RNM was extended with a control module that activates and inhibits muscles around the ankle of the stance leg, based on changes in whole body center of mass velocity. The added module improves the ability of the RNM to replicate human ankle torque response in response to perturbations. This reflex-based neuromuscular model with whole body center of mass velocity feedback can reproduce gait kinetics of unperturbed and perturbed gait, and as such holds promise as a basis for advanced controllers of prosthetic and orthotic devices.

Improving stability of people wearing a lower extremity Wearable Exoskeleton (WE) is one of the biggest challenges in the field. The goal of this preliminary study was to improve balance recovery from perturbations in people with incomplete Spinal Cord Injury (SCI) assisted by a WE with specifically developed balance controller. The WE has actuated ankle and knee joints, which were controlled by using a body sway-based balance controller. Two test pilots participated in 5 training sessions, devoted to enhance the use of the robot, and in pre/post assessments. Their balance during quiet standing was perturbed through pushes in forward direction. The controller was effective in supporting balance recovery in both tests pilots as reflected by a smaller sway amplitude and recovery time when compared with a minimal impedance controller. Moreover, the training resulted in a further reduction of the sway amplitude and recovery time in one of the test pilots whereas it had not an additional beneficial effect for the other subject. In conclusion, the novel balance controller can effectively assist people with incomplete SCI in maintaining standing balance and a dedicated training has the potential to further improve balance.

We investigated the capabilities of a reflex-based neuromuscular controller with a knee and hip gait trainer worn by a subject with a complete spinal cord injury. With controller assistance, this subject was able to reach a walking speed of 1.0 m/s. Measured joint torques agreed reasonably well with those of healthy subjects. The controller was also robust, recovering from manual swing foot perturbations. These preliminary results are promising for future implementation of neuromuscular controllers on wearable prototypes for real-world walking conditions. IntroductionRobust and reliable controllers of gait assistive devices for patients with neurological disorders must balance healthy-like walking function with promotion of motor recovery. We take a biologically-inspired approach with a controller based on a neuromuscular model developed by Geyer [2]. This neuromuscular controller (NMC) generates walking by activating simulated muscle reflex loops based on gait state, combining muscle-tendon dynamics to produce lower-limb joint torques. While the full controller can reproduce torques at the ankle, knee, and hip, its modular structure permits use for any combination of these joints. No pre-defined movement pattern is needed, and few sensors are required. Ground contact detection is needed to switch between stance and swing reflexes, and joint angles are used to calculate simulated

A neuromuscular model (NMC) presented by H. Geyer and extended by S. Song shows very interesting similarities with real human locomotion. The model uses a combination of reflex loops to generate stable locomotion and is able to cope with external disturbances and adapt to different conditions. However, to our knowledge no works exist on the capability of the model to handle sensory noise. In this paper, we present a method for designing Central Pattern Generators (CPG) as feedback predictors, which can be used to handle large amount of sensory noise. We show that the whole system (NMC + CPG) is able to cope with a very large amount of noise, much larger than what the original system (NMC) could handle.