Research On Motion Planning Of Reconfigurable Cable-Driven Parallel Robots

Cable-driven parallel robots(CDPRs)have parallel structures similar to traditional rigid ones,but their moving platform is connected to a fixed frame by cables.Benefiting from the structure of the cable-driven form,CDPRs have the advantages of large workspace,high payload-to-weight,and convenient construction.However,the fixed configuration limits its performance.To address this,a new kind of reconfigurable CDPRs(RCDPRs)have been proposed recently,which have a better performance with changeable configurations.As a relatively novel robot,its motion planning still suffers from many challenges.Therefore,in this dissertation,the following four aspects will be investigated:1)To solve the motion coordination problem of the cable anchors and cable lengths under multivariable strong coupling constraints in reconfiguration,a constrained path planning method that is suitable for RCDPRs is proposed,which can avoid obstacles by coordinating the changes of cable anchor positions and cable lengths.Considering the multivariate strong coupling characteristics of cable anchor positions and cable lengths,the loop-closure constraints is derived,and then the constrained configuration space is parameterized by an implicit manifold.Since the manifold is a space that has locally Euclidean properties,a set of tangent spaces is used to approximate it.Combined with the Rapidly-exploring Random Trees algorithm,tangent spaces are incrementally constructed,and the branches of the trees are expanded in the tangent spaces under the wrench feasible detection and collision detection,which greatly speeds up the convergence to constrained configurations.Finally,simulations and experiments are carried out in our self-built prototype,and the results shows that the feasible collision-free path of cable anchors and cable lengths can be found in the presence of multiple obstacles in the workspace.2)To minimize the number of reconfiguration actions for the RCDPRs,an optimal reconfiguration planning method is proposed to schedule the sequence and number of all movable cable anchors,so that the reconfiguration action can be reduced to save energy.Based on a two-level optimization,the method is divided into static reconfiguration(SR)and dynamic reconfiguration(DR),and in DR,the case of the RCDPR with an infinite number of configurations is modeled as solutions to a set of underdetermined linear equations.By imposing L0-norm optimization on these solutions,the configuration sequence with the minimum number of reconfiguration actions is selected.Since the combinatorial optimization of the L0-norm is difficult to solve,the L0-norm optimization is transformed into L1-norm optimization,which reduces the difficulty of solving.Simulations and experiments shows that the method can discover which reconfiguration is required and further minimize the number of reconfiguration actions.3)To improve the observation angle of Five-hundred-meter Aperture Spherical Radio Telescope(FAST),a novel upgraded feed cabin mechanism(UFCM)based on RCDPRs is proposed.Through the optimization of its structural parameters and antidisturbance analysis,it is proved that the UFCM can increase the observation angle of FAST and has strong anti-interference ability.First,a UFCM with six cables is proposed,and then its maximum tilt angle is calculated within tension limits.Based on this,the structural dimension optimization of the UFCM is optimized.Since the stability of the UFCM with six cables is poor,the UFCM with nine cables is proposed.Then,the structural dimension optimization of the UFCM is optimized for reducing the cable tension range.A disturbance increment method is proposed to linearize the nonlinear offset of the moving platform position caused by disturbances.Through the above analysis,it is verified that UFCM with nine cables can meet the observation requirements and has strong anti-disturbance capability.4)To deal with the motion impact of the UFCM with heavy loads due to multiple nonlinear mappings,an optimal jerk trajectory planning method is proposed.Combining with the requirements of observations,the trajectory planning of the UFCM is completed within a specific time,so that the radio source can be tracked smoothly and accurately.Based on the tracking mode of FAST,the tracked radio source is modeled and analyzed with the influence of FAST observation mode and the earth rotation.Then the trajectory of the cable length in the UFCM is obtained with the tilt angle,azimuth angle and kinematic model of the UFCM.In this way,the paths of the starting and stopping phase are limited to the feeder receiving surface and the cable tensions are kept within its ranges.Then,the jerk of the cable length trajectory is minimized with a septic B-spline,and the simulations show that the optimized cable length trajectory is smooth and continuous,which can not only accurately track the radio source,but also effectively reduce the impact generated by the trajectory.