Motion Planning And Control Of Cable-driven Segmented Linkage Manipulators For Narrow Space Operations

With the rapid development of aviation,aerospace,and nuclear power,the routine maintenance of large spacecraft,aircraft,and nuclear plants has become increasingly prominent.However,due to the narrow space and harsh environments such as extreme temperature and high radiation,the operation in such environment is at an extremely difficult level.Given that the cable-driven continuous robot has the characteristics of continuous and slim body,electromechanical separation,and super dexterity and adaptability,it has excellent potential for such applications.However,for its applications in above-mentioned fields,the problems of low structural stiffness,complex model,difficult motion planning and control are needed to be addressed in advance.With the problems aforementioned,this paper proposes a hybrid active and passive cable-driven segmented redundant manipulator with high stiffness and load capacity for the narrow space operations,and carry out its new structural design,kinematics modeling,trajectory planning,and dynamic control.Firstly,a hybrid active and passive cable-driven segmented linkage structure is proposed to reslove the problem of low stiffness and weak load capacity of cabledriven continuous robots.Generally,it is because the elastic parts of structural support for most continuous robots that leads to their aforementioned disavantages.Therefore,instead of elastic supports,this dissertation adopts the linkage cables to realize the fullly conostriant of each link and linkage of all dual degree-of-freedom(DOF)joints within the segment.The improved design of this structure owns many advantages.On the one hand,it can avoid the elastic supports and improve its structural stiffness;on the other hand,the segment can be bent with "constant curvature" by the sub-joint linkage,which can greatly simplify the manipulator modelling.Therefore,the new structure proposed in this dissertation can not only effectively improve the model accuracy,but also improve its stiffness and load capacity.To avoid the singularity problem and improve the efficiency of both the inverse kinematics resolving and configuration planning of cable-driven segmented robots,a fast inverse kinematics resolving and trajectory planning method is proposed based on two-layer geometric iterations(TLGI).Due to the manipulator's large number of joints,the traditional pseudo-inverse Jacobian based inverse kinematics resolving and planning method requires a lot of matrix operations,which has problems of singularity and low efficiency.Aiming at the problems above,this dissertation proposes a geometric iterative method,which has the adantage of fast resolving without suffering the singularity problem.On the one hand,this method disassembles the end pose into a specific vector and its rotational angle.The former and the end position are combined as the targets for the inner geometric iteration,and the latter can be solved directly during the outer layer iteration.On the other hand,this method simplifies the cable-driven segmented redundant manipulator into a discrete articulated arm,so that it can be suitable for the FABRIK(Forward and Backward Reaching Inverse Kinematics)method to achieve efficient inverse kinematics resolving.The simulation results show that the method has fast inverse kinematics resolving and planning capabilities,and could even have good behavior near the singular points.Furthermore,to make full use of the redundancy of the cable-driven segmented linkage robot so as to adapt to the more complex task,another inverse kinematics resolving and planning method for the obstacle avoidance is proposed based on extended virtual joints(EVJ).Since strong coupling relationship between the end pose and configuration,and the strict environmental constraints of the cable-driven manipulator,this dissertation introduces the equivalent virtual joint,which,on the one hand,can realize fast regional solution searching;on the other hand,can release equation constraints by adding virtual DOF and increase effective solutions.Therefore,the proposed method in this dissertation employs lots of adavantages,by combining the joints of manipulator itself and the EVJs which is converted by the task environment.On the one hand,it can realease the solution constraints for the end of the manipulator according to the given range of narrow working space;On the other hand,certain constraints can also be added to the local configuration to prevent the collisions.The simulation reuslts of typical tasks verify the effective of the method for the obstacle avoidance planning of manipualtor.The more difficult problems may exist in the dynamic control of such cabledriven segmented robots since its strong coupling,high nonlinearity and redundant DOF.Then,an equivalent dynamic model for the cable system is derived,and a PD control strategy with dynamic feedforward is proposed in this dissertation.Considering large number of driving and linkage cables of the robot,the dissertation firstly simplified the force analysis of driving cables,linkage cables,dual-DOF joint and links.In view of high efficiency of the Newton-Euler method in dynamic modeling,the recursive dynamics equation of the robot is derived.Combining the structural characteristics of in-segment equal-angle,and taking the minimum cable tension as the optimization goal,the optimal cable tensions of segment are obtained,and then the dynamic model of the entire robot is established.Furthermore,a PD control method is proposed based on dynamic feedforward to realize the closed-loop dynamic control of the robot.The co-simulation results using Adams and Matlab,varifies the accuracy of dynamic modeling and the effectiveness of the proposed control strategy.Finally,to verify the improved stiffness and load capacity of the proposed cabledriven segmented linkage robot,the prototype system is developed.Then the related experimental system is setup to carry out the performance experiments,including the stiffness,load capacity,linkage accuracy of the robot and etc.Based on the performance-based verification,the dynamic control experiment is further performed.Moreover,two other typical obstacle avoidance experiments are also carried out to verify the trjectory planning algorithms in this dissertation.The experimental results verify the stiffness and load capacity of the hybrid active and passive cable-driven linkage robot proposed in this dissertation,as well as the effectiveness and applicability of the related planning and control algorithms.