Project Overview
This project develops a novel framework for swarm robotics by integrating Reinforcement Learning (RL) with Linear Temporal Logic (LTL) to synthesize control policies for unmanned aerial vehicles (UAVs). The main objective is to enable swarms of UAVs to safely and efficiently navigate unknown environments through robust coordination, collision avoidance, and real-time adaptability.
Intellectual Merit
The research introduces a unique method of translating temporal logic specifications into RL reward functions, advancing the precision and reliability of swarm control. By combining distributed coordination strategies with dynamic policy optimization, the project addresses fundamental challenges in multi-agent systems—such as uncertainty, complex dynamics, and real-time responsiveness. The outcomes will contribute significantly to the specification, verification, and control of autonomous swarms, with broader applicability to other multi-agent domains.
Broader Impacts
Anchored at Alabama A&M University (an HBCU), the project empowers underrepresented students by engaging them in advanced AI, robotics, and machine learning research, inspiring future graduate studies and careers. Educational integration into computer science and engineering curricula provides hands-on learning opportunities, while the technology itself supports safer, more efficient drone operations in real-world applications. By broadening participation and fostering diversity in STEM, the project contributes to building a more inclusive innovation ecosystem.
Publications
Journal Articles
Wenzheng Fu, Yujian Fu, Jeanetter Jones, Zhijiang Dong, "Deep Learning Models for Microscopic Fungal Identification in Spacecraft Applications." Accepted by Advances in Artificial Intelligence and Machine Learning (ISSN: 2582-9793).
Mercy Akinyemi and Yujian Fu (2026). A study of embedded systems integrated with robotic platforms, demonstrating the transition from basic control to autonomous behavior. Embedded Systems and Robotics: From Microcontrollers to Intelligent Machines.
Asia Harris and Yujian Fu (2026). Programming quadruped robots: A survey of locomotion, control, and perception. International Journal of Computer Science and Mobile Computing, Vol. 15, No. 5, May 2026. doi: 10.47760/ijcsmc.
Conference Papers
Elton Mawire, Yujian Fu, Zhengtao Deng. Sensor-Driven Deep Reinforcement Learning with Hybrid Residual Safety Architecture for Autonomous Mobile Robot. International Conference on AI in Data Science and Robotics (ICAIDSR-26), Singapore, July 23, 2026.
Sharoon Sharif, Inikpi Egbunu, Yujian Fu, Zhigang Xiao. Hybrid Multi-Modal Learning for Extreme Long-Tail Fungi Classification. ACM Southeast (ACMSE), April 23-25, 2026, Troy, Alabama.
Thesis
Carlan Jackson. Machine Learning-based Drone Navigation and Sensor Mapping in Simulated 3D Environments. Thesis, Dec. 2025, Alabama A&M University.
AAMU STEM Day Posters
Team
Team Meeting (Date: Oct 3rd, 2025 — Time: 2:00 PM)