I am a Ph.D. candidate of Electrical and Computer Engineering in the joint doctoral program between University of California San Diego and San Diego State University, advised by Prof. Junfei Xie and Prof. Nikolay Atanasov. My research interests include Distributed Computing, Multi Agent Learning, Deep Learning and Unmanned Aerial Systems.
I received my M.S. degree in Computer Science from Texas A&M University Corpus Christi in 2019 and B.S. degree in Prospecting Technology and Engineering from Yangtze University, China in 2017.
[12.12.2022] One paper was accepted to the journal IEEE Transactions on Vehicular Technology.
[05.12.2022] I will start a summer internship at The Boeing Company as a Graduate Student Researcher.
[11.19.2021] I passed the University Qualifying Exam and became a Ph.D. candidate.
[07.10.2021] One paper was accepted to the journal IEEE Transactions on Network Science and Engineering.
DARL1N: Distributed multi-Agent Reinforcement Learning with One-hop Neighbors Baoqian Wang,
Nikolay A.Atanasov 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems
Abstract: In this paper, introduce a scalable MARL method called Distributed multi-Agent Reinforcement Learning with One-hop Neighbors (DARL1N). DARL1N is an off-policy actor-critic method that breaks the curse of dimensionality by decoupling the global interactions among agents and restricting information exchanges to one-hop neighbors. Each agent optimizes its action value and policy functions over a one-hop neighborhood, significantly reducing the learning complexity, yet maintaining expressiveness by training with varying numbers and states of neighbors. This structure allows us to formulate a distributed learning framework to further speed up the training procedure.
On Batch-Processing Based Coded Computing for Heterogeneous Distributed Computing Systems Baoqian Wang,
Shengli Fu IEEE Transcations on Network Science and Engineering, 2021
Abstract: Coded distributed computing (CDC) can efficiently facilitate many delay-sensitive computation tasks against unexpected latencies in distributed computing systems. In this paper, we focus on practical computing systems with heterogeneous computing resources, and design a novel CDC approach, called batch-processing based coded computing (BPCC), which exploits the fact that every computing node can obtain some coded results before it completes the whole task. To this end, we first describe the main idea of the BPCC framework, and then formulate an optimization problem for BPCC to minimize the task completion time by configuring the computation load.
Coding for Distributed Multi-Agent Reinforcement Learning Baoqian Wang,
2021 International Conference on Robotics and Automation (ICRA)
Abstract: This paper aims to mitigate straggler effects in synchronous distributed learning for multi-agent reinforcement learning (MARL) problems. Stragglers arise frequently in a distributed learning system, due to the existence of various system disturbances such as slow-downs or failures of compute nodes and communication bottlenecks. To resolve this issue, we propose a coded distributed learning framework, which speeds up the training of MARL algorithms in the presence of stragglers, while maintaining the same accuracy as the centralized approach. As an illustration, a coded distributed version of the multi-agent deep deterministic policy gradient(MADDPG) algorithm is developed and evaluated. Different coding schemes, including maximum distance separable (MDS)code, random sparse code, replication-based code, and regular low density parity check (LDPC) code are also investigated. Simulations in several multi-robot problems demonstrate the promising performance of the proposed framework.
Geolocation using video sensor measurements
Abstract: In this project, a simple geolocation algorithm using video sensor measurements is implemented in real robot. In particular, the mobile robot positions
in the image frame are obtained through YoloV3. Given the position of the camera in real world and intrinsic parameters of the camera, the position of the mobile robot in the real world frame can be obtained
through camera projection model. The estimated position are then compared to the real positions recorded by motion capture system.
In project, we developed an Andriod App called Pocket Planner, which aims to manage the time efficiently by allowing users to add their daily events and schedules in a flexiable and easy way. The code is written in Android Java.