Robot Control


In a near future, many kinds of robots will support out daily lives. In such a situation, robots should changes its behaviors to satisfy their requirements in response to changes in physical and system environments. Therefore, we believe robot software development become important. Currently, this group target following two topics related to robot software development process.

  • Optimizing robot behavior by evolutionary computation
  • Self-adaptive software framework for robot control


Research Topics

Self-adaptive software framework for robot control

Every moment changes the state of physical environment in which robots work. The state of robot itself also changes (e.g. breakdown of hardware, depletion of battery, etc...). In case the system of the robots fail to satisfy their requirements due to the changes mentioned above, developers need to stop the system temporarily and have to update the system's software. However, it often cost a lot to stop the systems. Therefore, it is needed that the ability of the system to modify behaviors/configurations of the system itself to satisfy requirements. This ability is called “self-adaptability”.

In our research, we develop a framework for robot control software which equips self-adaptability. This framework aims to support developing the self-adaptive control software by providing the requirement/environment/behavior models, and providing mechanisms for monitoring environment, determining adaptation plans, applying the plan at runtime.

Self-adaptive software framework for robot control

The structures of robots have become more and more complex these days. This makes it difficult for a practitioner to design the controller for robots. Evolutionary computation has been successfully applied to designing the controller, optimizing the motions of robots to achieve given tasks, even if the correct control was unknown. In our laboratory, we study to solve the problems which occurs when we apply evolutionary computation to the controller design. Our research goal is to make robots more familiar and more useful in our lives.


  • Ryuichi Takahashi (Specially Appointed Assistant Professor)

Contact Information

Ryuichi Takahashi:

Research Results

Selected Publications

  1. Shengbo Xu, Hirotaka Moriguch and Shinichi Honiden, Sample Efficiency Analysis of Neuroevolution Algorithms on a Quadruped Robot, 2013 IEEE Congress on Evolutionary Computation (CEC' 13), June 2013 (to appear)
  2. Xu Shengbo, Hirotaka Moriguchi,Shinichi Honiden, Efficient Neuroevolution for a Quadruped Robot, The Ninth International Conference on Simulated Evolution And Learning (SEAL 2012), December 2012