Accepted for Publication at ICCAIRO 2018

S- . T. Yau High School Science Award - Bronze Medalist

An algorithm that learns pertinent subspaces of the original high-dimensional distribution mapping

Original Paper: Deep neural network based subspace learning of robotic manipulator workspace mapping

The manipulator workspace mapping is an important problem in robotics and has attracted significant attention in the community. However, most of the pre-existing algorithms have expensive time complexity due to the reliance on sophisticated kinematic equations. To solve this problem, this paper introduces subspace learning (SL), a variant of subspace embedding, where a set of robot and scope parameters is mapped to the corresponding workspace by a deep neural network (DNN). Trained on a large dataset of around 6 × 10⁴ samples obtained from a MATLAB implementation of a classical method and sampling of designed uniform distributions, the experiments demonstrate that the embedding significantly reduces run-time from 5.23 × 10³ s of traditional discretization method to 0.224 s, with high accuracies (average F-measure is 0.9665 with batch gradient descent and resilient backpropagation).