Real-world deformable/compliant object manipulation is challenging for robots. More works are currently paying attention to this area. This project will mainly investigate and design robotic manipulation solutions for deformable/compliant objects such as meats, fruits, fabrics, etc. Some exemplary research questions include:
- how to evaluate the performance of robotic manipulation solutions in more effective and accessible ways,
- how learning methods (such as deep learning and reinforcement learning) can help achieve better manipulation performance for deformable/compliant objects with tactile sensing,
- and how tactile sensing can be used for reactive manipulation of soft objects.