This article is outdated but I keep it as a reference. Thank you for your cooperation.

I am now looking for the next position after getting my Ph.D. from Tokyo Institute of Technology, Japan (expected: Mar. 2023).

[Japanese version]

I want to get a research job including postdocs, but not limited to. At this stage, I am broadly considering both academia and industry. So far, I have devoted studying how to control multiple moving agents for almost five years. I feel my skills can adapt to various fields.

If interested, feel free to contact me.

Contact

okumura.k [at] coord.c.titech.ac.jp / LinkedIn

Below, I focus on the content not overwrapping with my portfolio.

Table of contents:

  • What I did
  • What I want to do
  • Career
  • Skills
  • Desired conditions
  • Others

What I did

I am widely interested in controlling multiple moving agents, using techniques from a variety of fields such as planning, search, automation, combinatorial optimization, robotics, multi-agent systems, and distributed computing. My papers have been published at selective conferences on AI and robotics such as IJCAI-19, AAAI-21, ICRA-21, IROS-21, AAMAS-22, and ICAPS-22.

I am particularly interested in the following three topics.

1. Quick, salable, and adaptive multi-agent/robot planning

I have developed techniques that solve path/motion planning for multiple agents, enabling efficient real-time planning for hundreds or more agents.

  • PIBT: sub-optimal algorithm to solve multi-agent pathfinding iteratively
  • mapf-IR: iterative refinement framework for real-time path planning
  • CTRM: data-driven roadmap construction for multiple agents in continuous spaces
  • TSWAP: sub-optimal algorithm that simultaneously performs target assignment and path planning
  • SSSP: domain-independent multi-robot motion planning

2. Robust execution strategies for reality gaps

Robot execution of planning is a non-trivial issue in practice.

  • Causal-PIBT, TSWAP, OTIMAPP: integrated planning & execution schemes that address timing uncertainties such as delays
  • AFADA: multi-robot platform that engages online/lifelong planning and execution

3. Applications of multiple moving agents

The above two are foundations but I am also interested in applications.

The next figure well summarizes my research interest.

What I want to do

I wish to continue developing cutting-edge technologies to control multiple moving agents.

Specific research topics interesting to me

I list up promising but under-explored topics below.

  • Sampling-based multi-robot motion planning. I feel uniting techniques of sampling-based motion planning developed for single-robot and multi-agent pathfinding is a promising direction like SSSP though this is still in the early stage.
  • Incorporating machine learning as heuristics to multi-agent/robot planning like CTRM. I am not saying about pure ML-based behavior planning, rather, I am interested in integrating classical planning methods with ML.
  • Deepen theoretical foundations from views of distributed systems. Systems of multiple moving agents are inherently distributed so the execution of such systems is not trivial at all. I studied timing uncertainties in OTIMAPP but am also interested in fault tolerance, uncertainties about locations, etc.
  • Swarm interaction. I am interested in coordination/cooperation between “powerful” agents, rather than “weak” robots usually assumed in studies on swarm robotics.
  • Killer applications.

Career

I have been supervised by Prof. Xavier Defago since I was an undergraduate. In addition, I fortunately had opportunities to work at:

Please check my portfolio for details. CV is also available on the website.

Skills

Programming

The research repositories are available on GitHub. I often use C++ / Python / Julia / JavaScript.

English

Not so fluent but I have promoted my research in English (note: my supervisor is not native to Japanese.)

Notes

  • I like creating videos.
  • Got several funds so far.
  • Hardware development is outside of my abilities.

Desired conditions

  • Make achievements public with credit.
  • Salary: I will at least refer to my total income when I was a 2nd-year Ph.D. student.
  • Location: Anywhere on earth. I do not have a specific preference.

Others

  • I will stay in Paris from the end of March until the end of August 2022. So it is relatively easy to visit places in Europe.
  • The book list that I often use may help the matching.
    • “Introduction to distributed algorithms.” Tel, G. 2000.
    • “Multiagent systems: Algorithmic, game-theoretic, and logical foundations.” Shoham, Y., & Leyton-Brown, K. 2008.
    • “Automated planning and acting.” Ghallab, M., Nau, D., & Traverso, P. 2016.
    • “Introduction to algorithms.” Cormen, T. H., Leiserson, C. E., Rivest, R. L., & Stein, C. 2009.
    • “Planning algorithms.” LaValle, S. M. 2006.
    • “Principles of robot motion: theory, algorithms, and implementations.” Choset, H., Lynch, K. M., Hutchinson, S., Kantor, G. A., & Burgard, W. 2005.

Thank you for reading this article. Again, feel free to contact me if interested.