Prior to joining the CoMMA Lab, I completed my B.S. in Computer Engineering at the University of Pittsburgh and my M.S. in Electrical and Computer Engineering at Carnegie Mellon University. During my time at CMU, I worked in the Search-based Planning Lab from The Robotics Institute, advised by Professor Maxim Likhachev.
My research interests are in developing real-time robotic algorithms leveraging high-performance computing techniques. I am also interested in the decision and planning problems in robotics that require reasoning about physics and dynamics in the environment. I believe that providing solutions to these challenges is essential for robots to perform dexterous tasks in unstructured settings, such as domestic environments.
Actor-Critic models are a class of model-free deep reinforcement learning (RL) algorithms that have demonstrated effectiveness across various robot learning tasks. While considerable research has focused on improving training stability and data sampling efficiency, most deployment strategies have remained relatively simplistic, typically relying on direct actor policy rollouts. In contrast, we propose PACHS (Parallel Actor-Critic Heuristic Search), an efficient parallel best-first search algorithm for inference that leverages both components of the actor-critic architecture: the actor network generates actions, while the critic network provides cost-to-go estimates to guide the search. Two levels of parallelism are employed within the search – actions and cost-to-go estimates are generated in batches by the actor and critic networks respectively, and graph expansion is distributed across multiple threads. We demonstrate the effectiveness of our approach in robotic manipulation tasks, including collision-free motion planning and contact-rich interactions such as non-prehensile pushing.
@inproceedings{yang2026pachs,author={Yang, Hanlan and Mishani, Itamar and Pivetti, Luca and Kingston, Zachary and Likhachev, Maxim},booktitle={IEEE International Conference on Robotics and Automation},doi={10.1109/ICRA57385.2026.11697579},pages={14441--14448},title={Parallel Heuristic Search as Inference for Actor-Critic Reinforcement Learning Models},year={2026}}
arXiv
Sampling-based Kinodynamic Planning with World Models
Visual world models predict the consequences of actions from images, but planning with them is typically done by optimizers such as CEM and MPPI. However, these optimizers stall on long-horizon tasks where progress requires motion the objective does not reward, such as moving around a block before pushing it. This paper revisits kinodynamic sampling-based motion planning for world models, using the learned model as the propagator of a tree search in latent space. Our key insight is that the sampling strategies devised for high-dimensional search in configuration space apply directly to the latent space of a world model, where they rapidly expand through intermediate states that myopic optimizers never sample. We present two exploration strategies for the high-dimensional latent space, nearest-neighbor-guided RRT and projection-based KPIECE, with GPU-native batched expansion for efficient device utilization. We evaluate with an off-the-shelf DINO-WM on four simulated domains and a real-world block-pushing task. Our approaches match or exceed the success rate of CEM and MPPI at up to 13.8x lower planning time in simulation, and improve the success rate by 20% on the real robot.
@misc{gao2026wmplan,archiveprefix={arXiv},author={Gao, Yitian and Yang, Hanlan and Kingston, Zachary},eprint={2610.12910},note={Under Review},primaryclass={cs.RO},title={Sampling-based Kinodynamic Planning with World Models},year={2026}}
