Research Statement
My research is motivated by the fact that both Construction and Agriculture are the fields yet to harness the power of the integration of Automation, Robotics and Artificial Intelligence. Both industries are still at the nascent stage of deploying autonomous systems to ensure higher productivity, safety and precision of task executions. I am particularly interested in generating intelligent solutions to motion and/or manipulation planning and control problems of potential robotics and autonomous systems employed in the aforementioned industries.
I strongly feel that the best way to pave the path for Intelligent Robotics and Autonomous Systems is to bring together design, mathematical modeling, optimization and learning. My aspiration is to conduct research towards building a Sustainable Autonomous Planning and Control Ecosystem, leveraging digitalization, robotics, and computational intelligence, via the amalgamation of knowledge-driven and data-driven approaches. The sequential process - mechanical design > mathematical modeling of dynamical systems > optimal planning > learning-based control - can be realized by combining the following characteristics of model-based and model-free approaches:
I strongly feel that the best way to pave the path for Intelligent Robotics and Autonomous Systems is to bring together design, mathematical modeling, optimization and learning. My aspiration is to conduct research towards building a Sustainable Autonomous Planning and Control Ecosystem, leveraging digitalization, robotics, and computational intelligence, via the amalgamation of knowledge-driven and data-driven approaches. The sequential process - mechanical design > mathematical modeling of dynamical systems > optimal planning > learning-based control - can be realized by combining the following characteristics of model-based and model-free approaches:
- The ability of model-based algorithms to address causality in nonlinear and underactuated dynamical systems
- The Pareto-optimal abilities of evolutionary algorithms for multi-objective optimization problems of design, planning and control
- The task-specific latent space abstraction abilities of neural systems trained with reinforcement learning algorithms
- The guarantee of higher productivity and safety of robotics and autonomous systems
Current Projects
Planning Module for Robotic Prefabrication of Mass Timber Building Components
This research aims to investigate the robotization of the manufacturing stage of CLT panelized building components such as walls and floors to identify a potential optimal prefabrication system by establishing a seamless design-to-fabrication framework.
Design Module for Automated Prefabrication of Mass Timber Building Components
This research seeks to automate the design stage of CLT panelized building components to solve the tasks of optimal design for manufacturing and assembly. The primary focus is on the prefabrication of planar building components such as walls and floors.
