I am a roboticist. I am currently a postdoctoral research fellow working with Prof. David Hsu at the Department of Computer Science, National University of Singapore after receiving my PhD degree from the Nanyang Technological University. I have been focusing on tackling large-scale decision making problems in robotics that involve complex environments, uncertainties and long-term planning. My research interests include robot motion planning, decision making, robot learning, parallel computing, and their applications to autonomous driving in crowded environments. My goal is to enable robots to seamlessly interact with humans in crowded, chaotic environments and accomplish complex tasks. Please see this video for a 3-min introduction of my recent research, or check out my research statement and CV for details.
PhD in Robotics, 2016
Nanyang Technological University, Singapore
Bsc in Computational Mathematics, 2011
ChuKoChen Honors College, Zhejiang University, China
Research topics:
Research topics:
Research topics:
Specifications:
This paper presents GAMMA, a general motion prediction model that enables large-scale real-time simulation and planning for autonomous driving. GAMMA models heterogeneous, interactive traffic agents that operate under diverse road conditions, with various geometric and kinematic constraints. GAMMA treats the prediction task as constrained optimization in traffic agents’ velocity space. The objective is to optimize an agent’s driving performance, while obeying all the constraints resulting from the agent’s kinematics, collision avoidance with other agents, and the environmental context. Further, GAMMA explicitly conditions the prediction on human behavioral states as parameters of the optimization model, in order to account for versatile human behaviors. We evaluated GAMMA on a set of real-world benchmark datasets. The results show that GAMMA achieves high prediction accuracy on both homogeneous and heterogeneous traffic datasets, with sub-millisecond execution time. Further, the computational efficiency and the flexibility of GAMMA enable (i) simulation of mixed urban traffic at many locations worldwide and (ii) planning for autonomous driving in dense traffic with uncertain driver behaviors, both in real-time. The open-source code of GAMMA is available online.
When robots operate in the real world, they need to handle uncertainties in sensing, acting, and the environment dynamics. Many tasks also require reasoning about long-term consequences of robot decisions. The partially observable Markov decision process (POMDP) offers a principled approach for planning under uncertainty. However, its computational complexity grows exponentially with the planning horizon. We propose to use temporally-extended macro-actions to cut down the effective planning horizon and thus the exponential factor of the complexity. We propose Macro-Action Generator-Critic (MAGIC), an algorithm that learns a macro-action generator using feedback from a planner, and in turn uses the learned macro-actions to condition long-horizon planning. Importantly, the generator is learned to directly maximize the down-stream planning performance. We evaluate MAGIC on several long-term planning tasks, showing that it significantly outperforms planning using primitive actions and hand-crafted macro-actions in both simulation and on a real robot.
Autonomous driving in an unregulated urban crowd is an outstanding challenge, especially, in the presence of many aggressive, high-speed traffic participants. This paper presents SUMMIT, a high-fidelity simulator that facilitates the development and testing of crowd-driving algorithms. By leveraging the open-source OpenStreetMap map database and a heterogeneous multi-agent motion prediction model developed in our earlier work, SUMMIT simulates dense, unregulated urban traffic for heterogeneous agents at any worldwide loca- tions that OpenStreetMap supports. SUMMIT is built as an extension of CARLA and inherits from it the physical and visual realism for autonomous driving simulation. SUMMIT supports a wide range of applications, including perception, vehicle control and planning, end-to-end learning. We provide a context-aware planner together with benchmark scenarios and show that SUMMIT generates complex, realistic traffic behaviors in challenging crowd-driving settings.
Autonomous driving in a crowded environment, e.g., a busy traffic intersection, is an unsolved challenge for robotics. The robot vehicle must contend with a dynamic and partially observable environment, noisy sensors, and many agents. A principled approach is to formalize it as a Partially Observable Markov Decision Process (POMDP) and solve it through online belief-tree search. To handle a large crowd and achieve realtime performance in this very challenging setting, we propose LeTS-Drive, which integrates online POMDP planning and deep learning. It consists of two phases. In the offline phase, we learn a policy and the corresponding value function by imitating the belief tree search. In the online phase, the learned policy and value function guide the belief tree search. LeTS-Drive leverages the robustness of planning and the runtime efficiency of learning to enhance the performance of both. Experimental results in simulation show that LeTS-Drive outperforms either planning or imitation learning alone and develops sophisticated driving skills.
Planning under uncertainty is critical for robust robot performance in uncertain, dynamic environments, but it incurs high computational cost. State-of-the-art online search algorithms, such as DESPOT, have vastly improved the computational efficiency of planning under uncertainty and made it a valuable tool for robotics in practice. This work takes one step further by leveraging both CPU and GPU parallelization in order to achieve near real-time online planning performance for complex tasks with large state, action, and observation spaces. Specifically, we propose Hybrid Parallel DESPOT (HyPDESPOT), a massively parallel online planning algorithm that integrates CPU and GPU parallelism in a multi-level scheme. It performs parallel DESPOT tree search by simultaneously traversing multiple independent paths using multi-core CPUs and performs parallel Monte-Carlo simulations at the leaf nodes of the search tree using GPUs. Experimental results show that HyPDESPOT speeds up online planning by up to several hundred times, compared with the original DESPOT algorithm, in several challenging robotic tasks in simulation.
Cooperative dual-crane lifting is an challenging and critical task in industrial sites. In this paper, we aim to automatically generate optimized dual-crane lifting paths under highly complex constraints, i.e., collision avoidance, coordination between the two cranes, and balance of the lifting target. We propose a mathematical modeling of the cooperative lifting system. Based on the formulation, we devleop a massively parallel solver based on a multi-objective Genetic Algorithm to compute highly-optimized lifting trajectories that satisfy continous collision-avoidance, coordination, and load-balancing constraints in complex industrial envirnoments. Our results show that the planner generate lifting paths that are safe, efficient, and easy for conduction for any complex environments.