Research On System Design And Position Control Of Inflatable Robotic Arm

Robots effectively lower the life risk of humans who operate in dangerous enviroments,and have become a relatively effiecient and safe inspection tool.Among them,the inflatable robotic arms(IRAs)that own the merits of light weight,portability and low cost are able to address the the issue of limited environmental compliance of existing inspection devices,and have great application potential in the future inspection scenerios such as disasters and archaeology.Europe,America and Japan have become the representive to take the lead in conducting the research of IRAs.Whereas,the related IRA research is still in its nascent stage,and there still remain extensive issues to be addressed urgently.To accomplish the inspection tasks,this thesis develops the system prototype of an IRA that has the advantages of “lightweight and handy structure,foldable ability,safety and low cost”.Concurently,the model establishment and motion control theory of the IRA are studied in depth based on the prototype,and the effectivenss of the design,modeling and control methods is experimentally demonstrated.The design scheme of the IRA’s modular system is proposed,and the actuation manner,preesure control method,position sensing method and material molding process are also designed in detail,implementing the primary trade-off towards inspection applications;A pramatric design model that owns the capability of describing the charactiristics quatitatively is proposed,including the stiffness model of irregular-shaped inflatable tubes through integrating the beam and membrane models,and the effectiveness and superiority of the design model is verified by comparative experiments.A further joint trade-off among the various characteristics is accomplished by the modelbased optimal deisgn;The system prototype of the modular joint is established based on the optimization results,and the efficacy of the method is validated via performance tests and analyses.The model construction theory of the IRA is studied in depth,thus realizing the accurate analysis of motion chatractiristics.A completely analytical model of pneumatic artificial muscles is proposed based on the principle of virtual work.The prediction of the nonlinear and hysterisis charactirics is implemented by considering the elastic energy and the frictional enegy loss,and the effectiveness of the model is demonsatrated through comparative experiments.Based on the motion modality analysis of the manipulator,the series rigid-body method and the piecewise constant curvature assumption are compared by using the open-loop calibration method,thereby realizing the kinematic description of the IRA.For the IRA’s thin-walled and incomplete continuum structure characteristics,the dynamic modeling approach that recusively uses the Cosserat rod theory is proposed;The forward and inverse solution methods of the dynamics are given,which adopts the implicit differential method to simplify the boundary value problem of partial derivative equations;The parameter identification of the model is conducted,and the efficacy and accuracy of the model is demonstrated by comparative experiements;The effectiveness and superiority of model are further verified by the comparative experiments regarding the model-based tracking control of the manipulator.Given the strong nonlinearities of the IRA model itself and the uncertainties induced by both internal and external strong disturbances,the position control strategy is studied.By comparing the existing controllers,the hybrid adaptieve disturbance rejection control(HADRC)method is proposed with the aim of “little model information,adpatability,disturbance rejection”;The model-free adaptive control(MFAC)and the active disturbance rejection control(ADRC)are integrated therough the fuzzy logic control(FLC),and the smooth transition of the two sub-controllers are concurrently assured.In order to enhace the tracking performance during the inspection procedure,the convergence-enhaced MFAC is proposed from the perspective of the global superlinear convergence of errors;The multi-variable decoupling method of the multi-input and multi-output(MIMO)ADRC controller,which addresses the disturbance issue induced by the inspection environment and the manipulator itself;The effectivenss of the position control method is demonstrated by the comparative simulations and experimens under the conditions of various disturbances and different types of traking signals.To thoroughly verify the effectiveness of the method poposed by this thesis,and to embody the application potential of the IRAs,the IRA system that owns orthogonal joints is established towards long-range inspection.The function and performance of the system are validated,and are compared to those of the exsiting congeneric and typical research;On this basis,the simulation experiments regarding the obstacle avoidance and passingthrough motions in disaster and treasurable relics/plant enviroments.The results reveal that the test indicators of the system meet the design requriments;The proposed design method is able to implement the trade-off towards inspection applications,and the modeling and control methods contribute to realizing the reliable inspection in complex enviroments,whist ensuring the safety of the operation process;Therefore,the design,modeling and control methods proposed by this thesis satify the demand of IRAs’ inspection and end-effector manipulation.