Papers
12
Total Citations
219
H-Index
9
About
Zifan Xu is a robotics researcher specializing in autonomous navigation, reinforcement learning, and adaptive planning for mobile robots. His work addresses one of the field's central challenges: enabling robots to navigate reliably across diverse, constrained, and dynamic real-world environments without exhaustive expert re-tuning. Xu is best known for the APPLR and APPL series of works (39–43 citations each), which introduced adaptive planner parameter learning — a paradigm that replaces hand-crafted navigation parameters with learned, environment-responsive ones, substantially reducing the burden on human experts. Complementing this, his benchmarking contributions have shaped how the community evaluates navigation systems: he has been a key figure in the BARN Challenge series at ICRA (2022–2024), providing standardized, highly constrained navigation benchmarks that have drawn broad community participation. His survey on deep reinforcement learning for autonomous navigation (40 citations) critically examines safety limitations and real-world deployment gaps, while DynaBARN extends evaluation to dynamic obstacle environments. More recently, Xu has expanded into legged locomotion in confined 3D spaces and curriculum learning for sim-to-real transfer. With over 215 cumulative citations across a focused body of work, Xu has established himself as a significant contributor to practical, learning-enabled robot navigation.
Research Focus
Key Achievements
Top Papers
- 1APPL: Adaptive Planner Parameter Learning43 citations · 2022
- 2Benchmarking Reinforcement Learning Techniques for Autonomous Navigation40 citations · 2023
- 3APPLR: Adaptive Planner Parameter Learning from Reinforcement39 citations · 2021
- 4
- 5
- 6
- 7DynaBARN: Benchmarking Metric Ground Navigation in Dynamic Environments13 citations · 2022
- 8
- 9
- 10Grounded Curriculum Learning3 citations · 2024