Predicting 3D Human Motion in Human-Robot Collaborative Tasks using 3D Deep Learning Methods
The field of Human-Robot Collaboration (HTC) involves collaborative processes where humans and robots work together to achieve a shared goal. Such tasks take advantage of the capabilities and strengths of humans and robots to perform the desired task in an optimal manner. As an example, such a collaboration can take advantage of human accuracy and ability to adapt to the work scenario, together with the robot’s capabilities; repeatability, accuracy, and greater strength, to name a few. Therefore, there is a large interest in improving the performance of HTCs, while ensuring the continued safety of the humans involved.
The goal of this research is to develop a network capable of recognizing and predicting 3D human motion using 3D deep learning methos and comparing its performance to the current state of art in the field. The impact of different network architectures will also be analyzed as well as the performance of the model when using synthetic images and real images.