​

• Title: Reinforcement Learning for Robotic Manipulation Based on Multi-Sensor Fusion

For sensing techniques used for robotic dexterous manipulation, there are two primary modalities: vision and tactile. The robot can be guided to approach the target objects based on visual feedback, while tactile feedback can provide useful information of interaction between objects and the environment. These two modalities are complementary to each other for contact-rich tasks.  To this end, this project aims to explore multimodal representation learning, which could be a useful tool for developing a robust and task-related representation to facilitate reinforcement learning with high efficiency.

​

• Title: Sim-to-Real Transfer Learning for Robotic Manipulation

Training a deep learning model for a robotic manipulator to locate and grasp the target object is notoriously costly and time-consuming. It requires a large amount of data, which is not reasonable for most real-world applications. This project aims to train the deep learning model in a simulation environment and implement a sim-to-real transfer learning technique to transfer the pre-trained model from the simulator to the physical environment.

​

• Title: Design of a two-finger robotic hand with tactile feedback

This project aims to develop a two-finger robotic hand, which can be used to grasp deformable target objects such as plastic bottles with various shapes. The robotic hand should integrate tactile sensors to provide feedback to the controller, which paves a way for the development of an adaptive control algorithm to determine the best grasping strategy.

​

• Title: Soft robotic hand for in-hand manipulation

This project aims to develop a soft robotic hand and demonstrate its potential for in-hand manipulation. In-hand manipulation involves manipulating an object within one hand. It is a vitally important capability for robots to accomplish real-world tasks. Therefore, it is worthwhile to investigate robotic dexterous in-hand manipulation to reconfigure an object. Soft robotic hands can enable robots to grasp delicate or soft objects of varying shape, size, and pose without explicit knowledge, and is promising for potential applications such as manipulating eggs or fruits etc. Machine learning approaches can be explored to mitigate the need of complex planning and control for in-hand manipulation when using a soft robotic hand with dexterous fingers.

​

• Title: Surgical Gesture Segmentation and Recognition for Medical Robotic Applications

Minimally invasive surgery mainly consists of the execution of a series of specific sub-tasks, which can be further decomposed into basic gestures or contexts. As a prerequisite of autonomic operation, surgical gesture recognition can assist motion planning and decision-making, construct context-aware knowledge to improve the surgical robot control quality, which may lead to better clinical outcomes. This project aims to investigate Bayesian Inference and explainable machine learning approaches for surgical gesture segmentation and recognition across multiple robotic platforms. The student is expected to have basic knowledge of machine learning and is familiar with Python.

​

• Title: Human-Robot Cooperative Control for Medical Robotics

Teleoperation is widely used for medical robotic platforms. To ensure seamless and intuitive control, an optimized leader-follower control model should be built to assist the operator in a cooperative manner. The operator can retain precise control when conducting delicate or complex manipulation, while an accelerated movement to the next distant target can be achieved. Moreover, haptic guidance should be included in the control framework. Learning from expert demonstration algorithms can be explored to obtain a pre-trained model, which can be used to provide the haptic guidance signals for instructing trainees during teleoperation or supervising novice surgeons to enhance the surgical operation skills. The student is expected to have basic knowledge of machine learning and is familiar with Python/C++ as well as simulation.

​

• Title: Robotic Skill Learning for Cooking

With advances in the field of robotic manipulation, machine learning and sensing, intelligent robots are expected to assist humans in kitchens and restaurants. This project aims to develop assistive robots for cooking, which are envisioned to replicate human skills in order to reduce the burden of the cooking process.