Institution profile

Air Force Institute of Technology

Academic institutionnorthamerica · us
Official website
Research library8linked papers
Opportunities0open roles
Selected work

Representative Papers

Improved Representation of Matrix Lie Group Operations through Tensor Notation

Jun 08, 2026

This work addresses the complexity of derivative computation and algebraic expressions on matrix Lie groups in state estimation, which has long hindered algorithmic understanding and implementation. For the first time, it introduces tensor notation together with Einstein summation convention into the differential calculus of matrix Lie groups, integrating concepts from differential geometry to establish a concise and unified mathematical formalism. This framework substantially enhances the clarity and readability of derivative derivations and algebraic manipulations, thereby facilitating more intuitive comprehension and efficient implementation of gradient-based estimation algorithms.

0 citationsRead paper

Parallelizing the branch-and-bound with isomorphism pruning algorithm for classifying orthogonal arrays

Apr 17, 2026

This work addresses the high computational complexity and limited scalability in enumerating non-OD-equivalence classes of large orthogonal arrays—specifically OA(128,9,2,4), OA(144,9,2,4), and OA(192,k,2,4) for k=9,10,11—by proposing a parallelized branch-and-bound algorithm. The method integrates Margot’s isomorphism pruning strategy with symmetry reduction techniques to drastically shrink the search space. It achieves, for the first time, a complete classification of OA(192,k,2,4) for k=9,10,11, and demonstrates near-linear speedup on OA(128,9,2,4) and OA(144,9,2,4). These results overcome the scalability limitations of serial approaches and confirm the algorithm’s efficiency and practicality in tackling large-scale orthogonal array enumeration problems.

0 citationsRead paper

Taxonomy and Trends in Reinforcement Learning for Robotics and Control Systems: A Structured Review

Oct 11, 2025

This paper addresses the persistent gap between theoretical advances in reinforcement learning (RL) and their practical deployment in robotics and control systems. To bridge this divide, we propose a structured taxonomy tailored to real-world robotic applications, grounded in the Markov decision process (MDP) framework and systematically incorporating mainstream deep RL algorithms—including DDPG, TD3, PPO, and SAC—across canonical domains such as motion control, dexterous manipulation, and multi-agent coordination. The taxonomy explicitly integrates training paradigms and deployment maturity metrics. Crucially, we identify recurring design patterns and evolutionary trends in high-dimensional continuous control tasks, thereby unifying theoretical insights with engineering constraints. Our framework advances reproducibility, transferability, and robustness in RL deployment on physical robots, offering both a methodological foundation and actionable guidelines for practitioners. (149 words)

0 citationsRead paper
Recent publications

Latest Papers

Improved Representation of Matrix Lie Group Operations through Tensor Notation

Jun 08, 2026

This work addresses the complexity of derivative computation and algebraic expressions on matrix Lie groups in state estimation, which has long hindered algorithmic understanding and implementation. For the first time, it introduces tensor notation together with Einstein summation convention into the differential calculus of matrix Lie groups, integrating concepts from differential geometry to establish a concise and unified mathematical formalism. This framework substantially enhances the clarity and readability of derivative derivations and algebraic manipulations, thereby facilitating more intuitive comprehension and efficient implementation of gradient-based estimation algorithms.

0 citationsRead paper

Parallelizing the branch-and-bound with isomorphism pruning algorithm for classifying orthogonal arrays

Apr 17, 2026

This work addresses the high computational complexity and limited scalability in enumerating non-OD-equivalence classes of large orthogonal arrays—specifically OA(128,9,2,4), OA(144,9,2,4), and OA(192,k,2,4) for k=9,10,11—by proposing a parallelized branch-and-bound algorithm. The method integrates Margot’s isomorphism pruning strategy with symmetry reduction techniques to drastically shrink the search space. It achieves, for the first time, a complete classification of OA(192,k,2,4) for k=9,10,11, and demonstrates near-linear speedup on OA(128,9,2,4) and OA(144,9,2,4). These results overcome the scalability limitations of serial approaches and confirm the algorithm’s efficiency and practicality in tackling large-scale orthogonal array enumeration problems.

0 citationsRead paper

Taxonomy and Trends in Reinforcement Learning for Robotics and Control Systems: A Structured Review

Oct 11, 2025

This paper addresses the persistent gap between theoretical advances in reinforcement learning (RL) and their practical deployment in robotics and control systems. To bridge this divide, we propose a structured taxonomy tailored to real-world robotic applications, grounded in the Markov decision process (MDP) framework and systematically incorporating mainstream deep RL algorithms—including DDPG, TD3, PPO, and SAC—across canonical domains such as motion control, dexterous manipulation, and multi-agent coordination. The taxonomy explicitly integrates training paradigms and deployment maturity metrics. Crucially, we identify recurring design patterns and evolutionary trends in high-dimensional continuous control tasks, thereby unifying theoretical insights with engineering constraints. Our framework advances reproducibility, transferability, and robustness in RL deployment on physical robots, offering both a methodological foundation and actionable guidelines for practitioners. (149 words)

0 citationsRead paper