Innovative Design And Modeling Of Continuum-arm Cluster For Dynamic Target Capture

Dynamic target capture is one of the issues to be solved in the field of robotics.The traditional grippers usually grasp the static targets in a relatively fixed position,which has the limitation of being difficult to adapt to the large impact collision and strong uncertainty.Considering the characteristics of large tolerance,strong robustness,and rigid-flexible coupling for capturing dynamic targets,this dissertation presents a novel design of continuum-arm cluster inspired by the motions of multi-tentacled creatures like the sea anemone.A set of theoretical methods including design,modeling and application in dynamic target capture process is established as follows:Design of rigid-flexible coupling continuum-arm cluster: Inspired by the motion of contraction muscle and tentacles of anemones,continuum arms with different driving types and a deployable base for installing the continuum arms are designed respectively.Among them,the active arm achieves shape adaptability through a central backbone and under-actuated tendons.The drive rod of the passive arm is fixed with the base constraint disk resulting in a passive bending under the geometric constraint.In addition,the deployable base,based on the Sarrus and the spatial 8-bar mechanism,can change the position and posture of the continuum arms,while driving the active arms.Overall,the continuum-arm cluster is driven by a single cylinder to synchronize the movement of 12 output links in different planes to expand and contract.Kinetostatics model of continuum arms: A virtual-trajectory equivalence method and a modified multi-link equivalence method are proposed respectively for flexible mechanisms with large deformation to solve the existing difficulties.The former method converts the deformation of the flexible mechanism to the trajectory of the rigid movement pair.Then,the screw theory is applied to the analysis of the flexible large deformation without establishing a large number of local frames.The latter method overcomes the drawback of large errors caused by the traditional model under large deformation by redefining the angle of adjacent links and modifying the configuration parameters in real time.In the statics,the geometric constraint equations are introduced for the first time,and an analytic model of the continuum arms with only the displacement of drive rod as the input is established.Analysis of collision interaction between dynamic targets and the continuum-arm cluster: To address the problem that the existing dynamic models of continuum arms cannot describe the real deformation of the flexible mechanism after instantaneous collision.A modified multi-link equivalence is used to decouple the motion of the masses by adding elastic-damping links between the discrete masses,and the original length and angle of the linkage are modified according to the current configuration.The dynamic model is obtained by Newton iteration.The dynamic characteristics of the continuum arms under the geometric constraint are described by introducing the drive tension optimization algorithm.Then,the collision detection method between the continuum-arm cluster and the dynamic targets is established,and the simulation model of the dynamic target capture process is obtained.This model reveals the collision deformation and energy dissipation process of the dynamic target and the large deformation multi-target flexible system.Dynamic target capture platform and its adaptability test: The continuum-arm cluster is installed at the end of the KUKA robot,and the dynamic target collision signal is detected by the acceleration sensor.Test the dynamic target capture success rate at different collision speeds and incidence angles,whose results showed that when the dynamic target speed was less than 4m/s and the normal vector of the upper platform of the arm was vertical or oblique,the average capture rate near the center of the arms is 90%.The robustness of the continuum-arm cluster was further evaluated after the partial removal of the arms.The results show that the dynamic target at a speed of 3m/s is still captured with more than 60% success rate after the removal of less than 4 arms.Under the same conditions,when the dynamic target was successfully captured in the simulation model,the test capture success rate is generally more than 70%.Therefore,the proposed model can be used to analyze and obtain an optimal capture strategy.Finally,the ability of the continuum-arm cluster to capture dynamic targets with random projections and static irregular objects are also demonstrated.This dissertation provides a new approach for capturing dynamic targets and conducts preliminary ground tests,which is expected to be applied in the fields of highspeed target sorting on production lines,underwater grasping and space debris cleaning in the future.The revealed motion-mechanics transfer mechanism and collision dynamic characteristics of the flexible continuum arm provide a theoretical basis for related research.