Junyue Jiang
Papers
2
Total Citations
3
H-Index
1
About
Junyue Jiang is a rising researcher at the intersection of autonomous driving, reinforcement learning, and efficient 3D perception. His work addresses two critical challenges in embodied AI: achieving precise vehicle control and enabling real-time, high-accuracy object detection under computational constraints. In his highly cited 2024 paper, Jiang pioneered a novel trajectory-tracking method that integrates Deep Deterministic Policy Gradient (DDPG) with Frenet coordinates. By transforming vehicle states from Cartesian to Frenet space, his approach allows a reinforcement learning agent to naturally follow curved paths, significantly improving lateral control stability—a key contribution for autonomous navigation. Building on this, his 2025 work tackles the "accuracy-efficiency paradox" in 3D object detection. He proposed an optimal Mixture of Experts (MoE) system that dynamically balances model precision with inference speed, directly addressing the safety-critical latency issues in autonomous vehicles and open-world robots. With over 3 citations already on these foundational papers, Jiang is establishing himself as a forward-thinking engineer who bridges theoretical control methods with practical, real-time system design—making his work essential reading for students developing next-generation autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2