Elbow Design, Servo Motor Selection and Control Implementation of the Agile Parallel Kinematic Manipulator

Abstract: The purpose of this work was to contribute to the creation of a prototype of a new type of robot called Agile Parallel Kinematic Manipulator or AgilePKM for short. Designing, building and controlling a new type of robot is a task which goes beyond the scope of any master thesis project, but there are subtasks which are more suitable to handle within the available time frame. A principle design will be handled, meaning no actual hardware will be available or constructed. There are three subtasks that are addressed as described below. The AgilePKM consists of many joints and links where one of the most crucial joints is the elbow joint. The first subtask was to construct this joint to efficiently meet the requirements and advantages of the new robot structure. The elbow joint is a very exposed joint as it exclusively actuates vertical movements, carrying not only the payload but the forearm structure as well. This was to be done while facilitating sufficient movement for the extensive working space and minimizing weight and cost for the proof-of-concept prototype. To ensure that these targets were fulfilled an adapted needs identification process was carried out followed by benchmarking and concept generation. From this the most promising concept was chosen and further developed to reach the most suitable design. The second subtask was to select servo motors. This task involved making a lot of assumptions and asking for specifications which was far from final due to the lack of a mechanical design. This was the reason to present a general recipe on how to select a servo motor, in addition to presenting the final selection for the prototype. The specifications as well as the motor and driver are subject to change when building a second prototype. This thesis will then form a good base of knowledge for such changes. The third task was to implement kinematic transformations for the robot, and to integrate those in an available automation software platform. The selected platform was TwinCAT3 by Beckhoff using the integrated CNC kernel provided by ISG Stuttgart. The kinematics was already developed in Python for easy prototyping and debugging. It was manually translated to C++ for performance and then imported to TwinCAT via a C++ module. The ISG kernel takes care of the trajectory planning. It was experienced that after several hours with the tool one notices the restrictions of the tools, such as the inability to dynamically limit the acceleration of each axis. The report forms a good basis for the design of many types of robots. In the case of the AgilePKM, the selections and implementations made in this thesis will be directly built upon and the prototype will be realized in the year following this thesis work. As mentioned, no hardware is available at the time of writing.