Heng Cao

Design frame of a leg exoskeleton for load-carrying augmentation

This paper presents a design frame of a leg ex-oskeleton. The performance of an exoskeleton was analyzed by studying a model of a linear coupled 1-DOF human-exoskeleton system. The results showed that a low-impedance, anthropomorphic and well attached mechanical structure is benefit for the control of exoskeleton, and a direct force feedback control could also be implemented to further reduce the impedance. Three kinds of applied sensor interface were compared. A leg exoskeleton called ELEBOT was developed for validating control algorithm. The ELEBOT included four main parts: mechanical structure, hydraulic actuator, sensor and control system and power. The function of such a load-carrying robot is to reduce the injuries associated with the back and lower limbs after a long heavy walk. Experiment results with a simple force controller confirmed that ELEBOT could efficiently assist people walking with 30 kg payload now.

Design and control of a hydraulic-actuated leg exoskeleton for load-carrying augmentation

For the research on control method of leg exoskeleton for carrying payload, a leg exoskeleton prototype called ELEBOT (ECUST Leg Exoskeleton roBOT) has been built. To increase the load-carrying ability of exoskeleton, pseudo-anthropomorphic mechanical structure, a kind of efficient, low-flow pump-valve hybrid controlled hydraulic actuator, custom sensing shoes and other electrical components for control have been developed. A knee torque estimate algorithm based on GRF (ground reaction force) was applied to ELEBOT via a hydraulic direct force feedback PD controller. Experiments indicated ELEBOT could support itself and 30 kg additional payload with above control method when it was walking at a speed up to 3.2 km/h while following the operators maneuvers. These results indicate the control method discussed in this work is valid.

Identifying hand-motion patterns via kernel discriminant analysis based dimension reduction and quadratic classifier

Mechanomyographic (MMG) signal for prosthetic control has been investigated in recent years and encouraging results in hand-motion patterns identification have been achieved. In this paper, only two accelerometer sensors were used to record the MMG signal in the forearm of fourteen able-bodied people. A kernel generalized discriminant analysis and three linear dimension reduction techniques were applied to reduce the feature dimensionality and improve the class seperability, and then the simple and commonly used quadratic classifier was implemented to identify the four hand-motion patterns. The experimental results have shown that the average identification rate reaches to a high accuracy of 95.12±3.83% by utilizing the three features extracted by kernel generalized discriminant analysis, where two wrist-related patterns are easier to identify while hand close is the most difficult one. It is concluded that two-channel MMG signal is sufficient for identifying the recorded four hand-motion patterns, which made MMG signal another prospective alternative in prosthetic hand control applications.

Unidirectional variable stiffness hydraulic actuator for load-carrying knee exoskeleton:

This article presents the design and experimental testing of a unidirectional variable stiffness hydraulic actuator for load-carrying knee exoskeleton. The proposed actuator is designed for mimicking the high-efficiency passive behavior of biological knee and providing actively assistance in locomotion. The adjustable passive compliance of exoskeletal knee is achieved through a variable ratio lever mechanism with linear elastic element. A compact customized electrohydraulic system is also designed to accommodate application demands. Preliminary experimental results show the prototype has good performances in terms of stiffness regulation and joint torque control. The actuator is also implemented in an exoskeleton knee joint, resulting in anticipant human-like passive compliance behavior.

A novel LMI based swing-up robust controller for a serial double inverted pendulum

This paper presents a robust controller based on the application of the linear matrix inequalities (LMI) to solve the swinging up and stabilization problem of a serial double inverted pendulum (SDIP), which is of complicate nonlinearities and limitations of actuator and sensors. The designed H-infinity robust controller was optimized to realize the required objectives and eliminate the negative effect of external disturbance. From the results of experiments in laboratory and numerical simulations, the proposed controller tests and verifies a satisfying performance in time responses, and its comparison with sliding mode controller confirms the advantages and efficiency of control performance in this pendulum problem.

Simulation Study on Neuromuscular Model-Inspired Control Strategy for Variable Stiffness Actuators

This study aims to seek a bionic control strategy for variable stiffness actuators and discuss the feasibility of this kind of actuators for the joint actuation on rehabilitation devices, exoskeletons, and prostheses. A robotic joint control schematic for variable stiffness actuators (VSA) inspired by the neuromechanical control model of biological joint is presented, based on the similarity analysis between neuromuscular system and VSA control system. By setting an equivalent short range of joint for specific posture, this control schematic changes the mechanical stiffness of joint when the control object shifts from posture control to motion control. Simulation results show that a robotic knee actuated by VSA can mimic the primary dynamic characteristics of a biological knee during a gait cycle using the presented joint stiffness control method.

A Method of Feeding Conditions Monitoring of Journal Bearing in Tandem Cold Mill

This paper presents a real-time hydrodynamic lubricant feeding conditions monitoring system for journal bearing in tandem cold mill to avoid online burning accident. A new kind of sensor connector has been developed to perform as a measuring section applied with nondestructive testing technology. The hydrodynamic lubricant feeding conditions in the tandem cold mill have been characterized. The monitoring results show that the hydrodynamic lubricant feeding conditions have been proved to be good indicators for the safety production. This real-time hydrodynamic lubricant feeding conditions monitoring system can be used as a long-term monitoring platform and provide an early warning for the tandem cold mill’s abnormal performance.

Dynamic Simulation of Passive Walker Based on Virtual Gravity Theory

Passive dynamic walking is a high efficiency way for bipedal robots to reduce energy cost. To study the nature of passive dynamic walking, we propose a new dynamic simulation method and build a virtual passive walker prototype based on virtual gravity theory using ADAMS. The results of the simulation are believable by comparing with the former research results, and the influences on gait stability of three critical physical parameters are obtained by using the parameterization analysis tools. These simulation results show that the suitable initial parameter is a critical factor for a simplest passive walkers stability on slope ground and the dynamic simulation is an available way to study the bipedal robots walking.