Naoki Yokoyama

About

Naoki Yokoyama is redefining how robots perceive, navigate, and act in the human world. His research lies at the intersection of embodied AI, semantic navigation, and mobile manipulation, with a focus on enabling robots to operate intelligently in unstructured, human-populated environments. Yokoyama introduced **Vision-Language Frontier Maps (VLFM)**, a zero-shot navigation approach that leverages semantic knowledge to explore unfamiliar spaces—garnering 93 citations since 2024. He also pioneered **Adaptive Skill Coordination (ASC)** for long-horizon mobile manipulation tasks like pick-and-place, and developed **ViNL**, which allows quadrupedal robots to step over obstacles while navigating. His work on evaluation frameworks is equally influential: he co-authored the widely cited **Principles and Guidelines for Evaluating Social Robot Navigation Algorithms** (59 citations) and proposed **Success Weighted by Completion Time (SCT)**, a dynamics-aware metric that improves upon standard navigation benchmarks. Yokoyama has also challenged conventional wisdom in sim-to-real transfer, showing that lower-fidelity simulation can yield better real-world navigation performance. Through these contributions—spanning over 300 combined citations—Yokoyama is shaping a future where robots navigate not just efficiently, but intelligently and socially.

Research Focus

Key Achievements

Top Papers

1. 1
   VLFM: Vision-Language Frontier Maps for Zero-Shot Semantic Navigation

   93 citations · 2024
2. 2
   Principles and Guidelines for Evaluating Social Robot Navigation Algorithms

   59 citations · 2024
3. 3
   Is Mapping Necessary for Realistic PointGoal Navigation?

   36 citations · 2022
4. 4
   ASC: Adaptive Skill Coordination for Robotic Mobile Manipulation

   36 citations · 2023
5. 5
   ViNL: Visual Navigation and Locomotion Over Obstacles

   28 citations · 2023
6. 6
   Success Weighted by Completion Time: A Dynamics-Aware Evaluation Criteria for Embodied Navigation

   25 citations · 2021
7. 7
   Principles and Guidelines for Evaluating Social Robot Navigation Algorithms

   15 citations · 2023
8. 8
   Rethinking Sim2Real: Lower Fidelity Simulation Leads to Higher Sim2Real Transfer in Navigation

   13 citations · 2022
9. 9
   System architecture for autonomous mobile manipulation of everyday objects in domestic environments

   11 citations · 2019
10. 10
    ASC: Adaptive Skill Coordination for Robotic Mobile Manipulation

    2 citations · 2023

Key Collaborators

Contact & Links