Previous studies on robotic rehabilitation have shown that subjects' active participation and effort involved in training can promote the performance of therapies. In order to improve voluntary participants during training, assist-as-needed (AAN) control strategies regulating assistance according conditions been developed. Unfortunately, heterogeneity patients' motor function capability task space is not taken into account implementation these controllers. this paper, a new scheme called...

Robotic-assisted rehabilitation of the upper limb following neurological injury can achieve best possible functional recovery when patients are engaged in therapy. However, implementation active training is still difficult as it's challenging to detect human motion intention online and impose corresponding robot control. This paper introduces a novel upper-limb robot, proposes sEMG-driven (sEMG: surface Electromyography) torque estimation model based on artificial neural networks (ANN). The...

The kinematic redundancy of human arm imposes challenges on joint space trajectory planning for upper-limb rehabilitation robot. This paper aims to investigate normal motion patterns in reaching and reach-to-grasp movements, obtain the unique solution a five-DOF exoskeleton. Firstly, six-camera optical tracking system was used capture participants' during goal-directed or movements. Secondly, statistical analysis executed explore characteristics swivel angle, which revealed that angle can be...

Lower limb exoskeletons offer support for patients suffering from mobility disorders due to injury, stroke, etc. But these devices are not used in day-to-day life and environments their limited human-computer interface perceive handle different terrains tasks. In this paper, we introduce a simple vision-based environment perception pipeline lower- obstacle crossing The proposed consists of three stages, namely, ground plane detection, estimating location dimensions, tracking. To reduce noisy...

Robot assisted rehabilitation training is promising to be an effective therapy for post stroke patients, while it's still a challenge robot control achieve the objective of "Assist As Needed" (AAN) This paper introduces new method implement AAN repetitive upper limb in three steps: (1) subject's motion frequency and phase are estimated adaptively using artificial central pattern generator (CPG) algorithm, (2) reference trajectory generated based on minimum-jerk principle patient's...

This study proposes a new method for modeling the complicated dynamics between motor neural signal (surface electromyography, sEMG) and corresponding joint torque of muscle contraction (sEMG-driven neuromusculoskeletal model), which has potential to be used rehabilitation robot control neuromuscular evaluation after stroke, etc. In this model, activation are built based on Hill-type many parameters determined using optimization methods, training samples sEMG, angles, torques acquired with...

This paper presents a 3-DOF compact rehabilitation robot, involving mechanical structure design, control system design and gravity compensation analysis. The robot can simultaneously provide assistance for pronation/supination(P/S), flexion/extension(F/E) adduction/abduction(A/A) joints training. P/S F/E are designed to be driven by cable transmission gain high backdrivability, an adjustment plate is adopted decrease the distance between rotation axis of joint human wrist robot. In addition,...

Rehabilitation robot is promising to be an effective therapy tool for post stroke patients, while the optimal robotic paradigm still unknown, which partly because there no standard method performing clinical validation study of rehabilitation robot. On other hand, robots are normally equipped with several kinds sensors, but few researches on analysing and fusing training information recorded by In this study, we introduce a novel upper limb force feedback ability virtual reality function,...

This paper is devoted to modeling and controlling a new upper-limb rehabilitation robot which has parallel structure. Genetic algorithm (GA) successfully applied in parameter identification based on dynamic analysis of the robot. For accurate identification, joint velocities accelerations are computed by Kalman filter. By taking non-linear characteristics frictions into account, unknown friction parameters differ depending directions motion. Compared with traditional least square estimation...

The rehabilitation training with active involvement and efforts of impaired subjects can enhance the outcome therapies. To promote patients' voluntary engagement, a variety assist as needed (AAN) controllers have been proposed for robot-aided therapy. However, impairment level is not taken into account in implementation those control schemes. In this study, novel AAN controller developed upper limb paradigm uses Gaussian radial basis function (RBF) network to learn model subjects' motor...

We developed a 3D end-effector type of upper limb assistive robot, named as Assistive Robotic Arm Extender (ARAE), that provides transparency movement and adaptive arm support control to achieve home-based therapy training in the real environment. The proposed system composes five degrees freedom, including three active motors two passive joints at module. core structure is based on parallel mechanism. kinematic dynamic modeling are illustrated detail. framework calculates compensated force...

Impedance control has been widely used in rehabilitation therapies. Especially the strategies with adaptability are proved to be beneficial patients. In this paper, we present an adaptive impedance controller for upper limb robots. Firstly, stiffness and damping of patients estimated by least square method. Then, a reward function is defined encourage patients' efforts limit tracking errors. To maximize function, optimal robot determined based on estimation stiffness. order validate...

The Assistive Robotic Arm Extender (ARAE) is an upper limb assistive and rehabilitation robot that belongs to the end-effector type, enabling it assist patients with movement disorders in three-dimensional space. However, problem of gravity compensation for human this type crucial, which directly affects deployment or field. This paper presents adaptive framework calculates compensated force based on estimated posture 3D First, we arm joint angles real-time without any wearable sensors, such...

For lower limb exoskeletons to be used for daily activities, it has understand the diverse nature of its operating terrain. Computer vision offers an excellent avenue terrain recognition such applications. In this paper, we propose a heuristic vision-based approach in exoskeletons, using line fitting features and active sensor fusion with exoskeleton encoders enhance performance. Our recognizes more classes than traditional methods, including obstacles gaps, enabling better traversal. We...

It has been proved that minimizing the condition number of observation matrix, which is calculated from robot dynamic model and associated exciting trajectories, very effective for improving identification accuracy robotic dynamics. A relative simple beneficial reduction number, hence, several simplification methods have proposed in literature. However, existed cannot be used to efficiently process structural errors, will inevitably cause inaccurate estimation Therefore, a novel method based...