Research
See full list of Publications below.
Current Research
I am doing Robotics research as a PhD student in the Machines in Motion Laboratory at New York University (NYU). My current interest is optimal control and reinforcement learning for exoskeletons and manipulators. More details will be added in the future.
Past Research
Below are some of my past research on optimal task and motion planning for mobile manipulators in the Control Robotics Intelligence (CRI) Group at Nanyang Technological University (NTU Singapore) from 2021-2024.
1. Mobile Manipulator Trajectory Optimization under End-effector Trajectory Continuity Constraint
- Applications: tasks requiring the end-effector to traverse a continuous (time-parametrized) trajectory, e.g. Mobile 3D Concrete Printing (developed at Singapore Centre for 3D Printing)
- Problem: find optimal base trajectory under constraints \(\begin{equation} \begin{split} &\mathbf{x}^{opt}(s) = \arg\min_{\mathbf{x}(s)} J[\mathbf{x}(s)]\\ \text{s.t.}\quad &\mathbf{x}(s) \in \mathcal{X}_a, \quad \dot{\mathbf{x}}(s) \in \mathcal{V}_a \quad \forall s \in [0,1] \end{split} \end{equation}\)
- Cost $J[\mathbf{x}(s)]$: integral of weighted squared velocity (which encourages steady motion, avoids abrupt changes)
- Constraints $\mathcal{X}_a, \mathcal{V}_a$: include end-effector trajectory continuity constraint, arm’s joint limits, base’s velocity limits and collision avoidance
- Method:
- Based on the required end-effector’s trajectory and other constraints, find the admissible configuration spacetime for the mobile base.
- Dynamic Programming to find the optimal base trajectory.
- Inverse Kinematics to compute the joint trajectory of the arm.
- Paper: Planning Optimal Trajectories for Mobile Manipulators under End-effector Trajectory Continuity Constraint (ICRA 2024) (PDF | Video)
2. Mobile Manipulator Task Sequencing
- Applications: tasks involving multiple targets distributed in a large workspace beyond the reachability of the robotic arm, e.g. drilling in a shopfloor (inspired by Airbus Shopfloor Challenge)
- Problem: find minimum number of base poses and shortest motion sequence to visit all targets
- Method:
- Task-space clustering: based on kinematic reachability analysis, cluster the task space into a minimum set of target clusters such that there exists a reachable base pose for each cluster to reach all inside targets.
- Task sequencing: compute the shortest base sequence, shortest target sequence, then shortest manipulator joint trajectory to visit all targets.
- Paper: Task-Space Clustering for Mobile Manipulator Task Sequencing (ICRA 2023) (PDF | Video)
Selected Publications
Full list: Google Scholar
2024:
- Q.-N. Nguyen, Q.-C. Pham, “Planning Optimal Trajectories for Mobile Manipulators under End-effector Trajectory Continuity Constraint,” International Conference on Robotics and Automation (ICRA 2024) (PDF | Video)
2023: