Mechanical Analysis And Performance Optimization Of The Cable-driven Parallel Robot

The cable-driven parallel robots(CDPRs) can offer high load/mass ratio, large workspace, rapid response and strong environment adaptability, which have abroad applications and valuable exploitations, play an important role in the family of parallel manipulators.In studies of CDPRs, workspace is the most basic sector and a prerequisite for the rest studies. There are two determinative factors of workspace: one is the “force” factor, ensuring CDPRs can balance external forces while keeping positive tensions, plays a main role in determining workspace; the other one is the “geometry” factor, guaranteeing each component not to interfere with other components, plays a secondary role in determining workspace. The essence of the “force” factor is the unilateral nature of cables, which leads to the actuation redundancy of CDPRs, causing the non-unique solutions to cable tensions. The unique solution to cable tensions can be determined by optimal distributions. When conducting the real-time force control on CDPRs with cable tension optimal distributions, it requires not only real-time calculation of tensions but also continuous tensions. However, cable mass and inertias are neglected when distributing cable tensions optimally for CDPRs in the past research, yielding the tension optimal distributions are not applicable to the long-span, actuation redundant and high-speed CDPRs(LRHCDPRs), which should be accounted in the dynamic modeling. Obviously, owing to the long-span, redundant actuations and high-speed motions, LRHCDPRs have the ability of large overall motions and high maneuverability, are a emerging development direction of CDPRs. Cable-driven parallel camera robot(camera robot for short) is a typical representative manipulator of LRHCDPRs. However, the performance of the camera robot is influenced seriously because of its high-speed motions; furthermore, the camera robot suffers from meteorological disturbance in operations. Therefore, a optimum structural design that gears to actual working characteristics of the camera robot is desiderated.This dissertation studies CDPRs from four key aspects, i.e.(1) analyze the workspace by combination of “force” and “geometry” factors;(2) evaluate the continuity andreal-time performances of the cable tension optimal distributions from the quantitative analysis perspective;(3) establish the complete dynamic models of CDPRs;(4) propose the multi-objective optimum structural design that gears to actual working characteristics of the camera robot. The relevant work is supported by the National Science Foundation of China(51175397). The research results lay a solid study foundation for performance improvement and engineering application of the cable-driven parallel robots. Furthermore the development bottleneck for the camera robot can be break. Thus, the cable-driven camera robot with the autonomous intellectual property will bring a breakthrough of camera boarding technology in China. The main works and creative achievements can be described as follows:The first part focuses on the interference regions within the wrench-feasible dynamic workspace. Three interference classifications of cable-driven parallel robots are clearly defined and then the causes corresponding to each kind of interference are analyzed. According to the clear interference classifications, the corresponding mathematical expressions of judging interference are obtained. Combined with the method of determining the wrench-feasible dynamic workspace, a complete, necessary and sufficient interference solution method is proposed to determine the interference regions within the wrench-feasible dynamic workspace. The research is helpful in deepening the recognition of the essential cause of existing interference and provides a valuable reference for the interference-free workspace.The second part focuses on the continuity and real-time performances of the cable tension optimal distributions. The definitions of continuity and real-time performances of the cable tension determining are defined. In order to evaluate the continuity and real-time performances, the indices are presented. Based on the “convex optimization theory”, “extract roots formula” and “polynomial extremum theorem”, a cable tension optimal distribution algorithm is proposed for the completely restrained cable-driven parallel robots, for which the continuity and real-time performances are proved. The simulation results show the differences and characteristics of continuity and real-time performances among the different optimization objects used for the cable tension optimal distribution algorithm. Furthermore, some valuable proposals of selecting the optimization objects are also offered.The third part focuses on the cable tension distributions considering cable mass and inertia force. The motions of the mass points of the cable are decomposed into two kinds of motions: i.e. the axial motions with the cable length variations in the vertical plane of the cable and the motions orthogonal to the vertical plane. Based on the two kinds of motions and ideas of the finite-element method, the dynamic model of the mass point in the cable is established. Then the complete dynamic model of the cable can be established by integration assembly operation. Furthermore, by using the Newton-Euler method, the complete dynamic equations including the cable dynamic are derived for cable driven parallel robots. To eliminate the uncertainty of the solutions to tensions caused by redundant actuations, a tension optimal distribution model is established for completely restrained and redundantly restrained cable-driven parallel robots. Based on the tension optimal distribution model, an iterative algorithm is presented to determine the tensions, which takes the cable tensions and lengths obtained by straight line model as the iterative initial values and the unchanged sags of cables as the terminating condition. Consequently, the optimal tensions are determined, which lay a theoretical foundation for force control to cable-driven parallel robots.The fourth part focuses on the multi-objective optimum structural design of the camera robot, which can improve the overall performance of the camera robot. The multi-objective structural optimization model of the cable-driven parallel camera robot is established combining with the ideal point and penalty function method, which synthesize the abilities of the tracking photography, anti-wind disturbance and tension impact action. Due to the weak capability of global optimization, an improved genetic algorithm with more powerful global optimization capability is presented,which has three improvement measures, i.e., the fitness calibration, sub-populations division and the adaptive crossover/mutation rates with the information entropy of the population varies(Adaptive Multi-island Genetic Algorithm based on Information Entropy of the Population,IEPAMGA). Then, the improved algorithm(IEPAMGA) is applied for solving the optimization problem. Under the condition of satisfying the upper-lower limit of tensions, minimum photography task-space and power cutoff frequency, the design parameters of the cable-driven parallel camera robot for the larger wrench-feasible dynamic workspace volume, higher global first-order natural frequency and smaller total tension impulse.