Optimization design and key-term separation identification of parallel six-dimensional pose sensor with high sensitivity and high precision

The integrated bearing-positioning parallel manipulator is an important mean for ground testing of space optical telescope before being launched into space. To ensure positioning accuracy, the manipulator requires a pose sensor for the end-effector measurement with six degrees of freedom (DOF), real-time, high-precision, and operation in a vacuum environment. However, it is challenging for the current pose sensor to simultaneously meet all these requirements. In this paper, a novel concept of six-dimensional pose sensor based on parallel mechanism is proposed. A new mechanism design approach is presented to achieve the large measurement range, real-time measurement performance, and high accuracy measurement based on performance atlas. First, the G_F sets synthesis is developed to optimize the configuration of pose sensor. Next, two new design indicators are proposed to evaluate the real-time and high accuracy performance. Mechanism parameters are optimized by combining with mechanism singularity analysis. To guarantee high measurement accuracy, a key-term separation identification method with double neural networks is presented. Experiments on 6-UPS pose sensor demonstrate the effectiveness of the proposed mechanism design and key-term identification approach.