Research On Hybrid-drive Full-body Powered Exoskeleton Robot

Along with scientific and technical development,growing importance has been attached to the concept of human enhancement.The application fields of human enhancement technology range from helping the seniors and the disabled to accomplishing the impossible load-bearing tasks for ordinary people,such as lifting and carrying large-mass cargos and executing special tasks while bearing large load in the long term.Exoskeleton robots are emerging in recent years as a wearable power assist robot,which combines human intellect with machine power.It can achieve a power output much large than human capacity without interfering with human body movement.In addition to provide extra power for the old,disabled or the normal people,exoskeleton robots have also found applications in earthquake rescue,fire rescue,hiking and space & aeronautics.Many domestic studies have been conducted in exoskeleton robots,and upper-and lower-limb exoskeleton robot prototypes have been developed.However,very few achievements have been made when it comes to full-body powered exoskeletons,which can walk on various road surfaces and lift and carry loads.Besides,the modeling of full-body power exoskeletons and research and development of the control system are still in its infancy.In this study,we develop a full-body power exoskeleton oriented towards lifting and carrying capacities.An innovative design is performed based on the motion features of the upper and lower limbs,so that the exoskeleton will possess greater load-carrying capacity.Moreover,an in-depth discussion is conducted over human movement recognition and control of full-body powered exoskeleton.Finally,the theory of exoskeleton is verified through a specific demonstrative experiment.The major difficulty encountered in the design of upper-limb exoskeleton is that a large load carrying capacity is hardly achievable with a long arm of the force.In the present study,a gas spring is used as a passive energy storage to achieve an upper-limb exoskeleton where an active mode is replaced by the passive mode.In addition,the torque problem of the motor is solved,and compared with the conventional spring structure,the use of gas spring balance also enhances the compactness of the structure.Then structural optimization is performed based on the output performance of the gas spring,so that the overall balance performance is improved.Finally,the control mode is optimized based on the adjustment time and capacity of the overall system.Our study lays the basis for the operation of the full-body powered exoskeleton.Lower-limb exoskeleton has a more complex mode of motion than the upper-limb exoskeleton.Firstly,we analyze human anatomy and perform simulations on human walking,squatting and rising,and climbing up and down the stairs.Range of motion and power consumed during these motions are obtained,based on which the lowerlimb exoskeleton is designed.Through the connecting rods and screws of the joints,variable reduction ratio transmission has bee achieved in the jionts,and the drive torque of the motor is reasonably allocated to the joints from different angles.Based on simulations and structural demands,the position and length of the connecting rods are optimized.Finally,a closed-loop control scheme for the lower-limb exoskeleton is developed which comprises of joint encoder,gyroscope,signals of plantar pressure and connecting rod force.For a linkage of the lower-and upper-limb systems,the plantar pressure feedback signals are collected and used for real-time recognition of human movement phase by fuzzy neural network.Thus important parameters are provided for the modeling and control of the linkage of the lower-and upper-limb systems.For the modeling of the upper-limb system,it is equivalent to a mass block with coordinates of moving centroid.The position vector of its center of gravity is calculated.The position vector of the center of gravity for the upper limb is used for kinetic modeling of the lower limb system.The human movement phase recognized by the fuzzy neural network is used for kinetic analysis,and kinetic models for different phases are obtained.Finally,based on the computing results of kinetic models and the interaction force feedback of the connecting rods,human-machine interaction control that integrates sensitivity amplification control feedforward and closed-loop control of force is achieved.Therefore,a stable power assist walking is realized for the full-body exoskeleton.To test for the power assist performance of the exoskeleton,an experimental platform consisting of the exoskeleton and the experimental site is established.The experiment is divided into three parts: test on upper-limb exoskeleton,test on lowerlimb exoskeleton and test on full-body powered exoskeleton.Firstly,the load-bearing capacity of the upper-limb exoskeleton is determined using the platform for upperlimb exoskeleton,and its power assist performance is evaluated.Next,the wearing comfort,kinematic flexibility and environmental adaptability are tested for the lowerlimb exoskeleton.The load-bearing capacity and speed capacity of the lower-limb exoskeleton are tested based on data of joint movement and human-machine interaction.Finally,the power assist performance of the full-body powered exoskeleton is evaluated by movement wearing the full-body powered exoskeleton by doing human-machine efficacy and physiology tests.