Research On Flexible Pneumatic Trunk-Type Continuum Robot

Robot technology is a combination of computer, control theory, mechanism, information and sensor technology, artificial intelligence, bionics and other high-tech, which is very active in contemporary research. And its applications are widespread increasingly. In addition to the traditional discrete robot such as industrial robot, continuum robots are attracting more attention of scholars at home and abroad for their excellent flexibility. The continuum robots are inspired elephant trunk, octopus arms which possess remarkable manipulation abilities without skeletal structure. The continuum robots seek to mimic the astonishing abilities of these organisms through designs which lack the rigid links which compose time-honored robots.For example, elephants use their trunk picking leaves from the tree, wrapping branches from the ground, and so on. By wrapping objects of widely various sizes and prior unknown shapes, continuum robots may utilize whole-arm grasping to manipulate objects too large, too fragile, or too unstructured for traditional robots. And they also could be used in search and rescue operations in the unstructured, confined environment. But compared to traditional industrial robots, continuum robots are not mature either in structural design or in the theoretical study. The flexible pneumatic actuator FPA was studied by our researching group. And based on it, the flexible pneumatic bending joints and flexible pneumatic ball joints were proposed. In this dissertation, the flexible pneumatic ball joint and the new flexible pneumatic bending joint are studied further. And based on them, a trunk-type continuum robot was proposed and studied.Firstly, the original design of the flexible pneumatic ball joint was improved, and its forward and inverse position expressions were studied based on geometry. The relationship between joint’s positions, orientations and the lengths of FPA in the joint was derived. Meanwhile, the original design of the flexible pneumatic bending joint was improved too. And its static model was simplified. A test platform was designed, on which various bending joints were tested. Based on the experimental data, an empirical model was derived which described the relationship between the joint length, output force and internal pressure. The proposed empirical model not only simplified, but also fixed the inadequacies of the original static model. Based on the research described above, and combined with the basic principles of trunk bionics, a flexible pneumatic trunk-type continuum robot was proposed in this dissertation. The proposed continuum robot was composed with flexible pneumatic bending joint and flexible pneumatic ball joint. Four ball joints were series connected to mimic the elephant trunk’s body, and two bending joints were applied to mimic the trunk’s finger synaptic. Pinching smaller objecs and wrapping bigger objects could be realized by the continuum robot. Three different kinds of kinematics were proposed based on the assuming that the tunk-type robot shape could be simplified as an arc. And the proposed kinematics expressions are suitable for other continuum robots.The flexible pneumatic ball joint statics equation was proposed based on the basic principle of screw theory. And the grip model of flexible pneumatic trunk-type continuum robot was studied by combining the trunk bionics and screw theory. In this dissertation, the modle of the continuum robot grabbing objects with various shapes was simplified as "three points arc grabbing". And its steady grab model and grab optimization algorithm was analyzed.Based on the research described above, the control system of the flexible pneumatic trunk-type continuum robot was implemented, which was composed with industrial computer (IPC) and embedded systems. The drive variables and robot orientation were translated by IPC, and the users’operating data was read by IPC. The neural network algorithm was applied in this system to translated drive variables and robot orientation. The sensor status and joint variable output were accomplished by the embedded systems.