Papers
9
Total Citations
159
H-Index
6
About
Javier Yu is a robotics researcher whose work sits at the intersection of distributed optimization, multi-robot systems, and neural scene representations. His most recognized contribution is DiNNO (2022, 37 citations), a pioneering algorithm enabling groups of robots to collaboratively train deep neural networks over mesh networks without sharing raw data—an elegant solution to privacy and bandwidth constraints in decentralized robotic teams. This work is complemented by his influential two-part series on distributed optimization for multi-robot systems (2024, 31 and 29 citations respectively), which serves as both a practical tutorial and a comprehensive survey, filling a notable gap in robotics literature. Yu has also made significant strides in robot navigation using modern 3D scene representations. His Splat-Nav framework (2025, 22 citations) introduced real-time safe navigation within Gaussian Splatting maps, while follow-on work such as SAFER-Splat and HAMMER extends these ideas into safety-critical filtering and heterogeneous multi-robot mapping. Earlier work on non-myopic task selection for heterogeneous teams (2018) reflects a longstanding interest in coordinated autonomy. Collectively, Yu's research advances the foundations of scalable, safe, and intelligent multi-robot collaboration—making his tutorials and algorithms increasingly essential resources for the robotics community.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Distributed Optimization Methods for Multi-Robot Systems: Part 2—A Survey29 citations · 2024
- 4Splat-Nav: Safe Real-Time Robot Navigation in Gaussian Splatting Maps22 citations · 2025
- 5
- 6A Survey of Distributed Optimization Methods for Multi-Robot Systems6 citations · 2021
- 7
- 8HAMMER: Heterogeneous, Multi-Robot Semantic Gaussian Splatting5 citations · 2025
- 9