September 2021 OES Beacon

Member Highlights (September 2021)

 Contact the editors if you have items of interest for the society

Research Activities in Underwater and Land Robotics

Takumi Matsuda, Meiji University, IEEE OES Japan Chapter, BEACON Associate Editor 

Introduction

Figure 4 The author’s photo taken together with AUV Tri-TON.

I have been researching a navigation method for multiple autonomous underwater vehicles (AUVs). The oceanographic survey by AUVs is generally conducted based on a ship. Due to limited survey opportunities and high operational costs, we need a survey method with minimal ship support. Therefore, I am researching a method to replace the ship with an AUV. The AUV, which is capable of accurately estimating its state, acts as a parent AUV which supports other AUVs with minimal sensors (child AUV). In this way, the parent and child AUVs behave as a single survey system and perform the survey [1]. Figure 1 shows the deployment scene of 3 AUVs in the sea experiments.

I moved to Meiji University (https://youtu.be/zjYBDoAcNrg) from the University of Tokyo last year, and have been studying with Professor Yoji Kuroda while applying the idea of parent-child robots to land robots to develop monitoring technology by robots for indoor environments such as offices, airports, and stations.

Underwater and Land Robotics

Cost is an important factor in underwater robots, but it is also important for land robots. Currently, indoor environment security, which is carried out by humans and surveillance cameras, is increasingly being carried out by robots. Normally, a single robot surveys in an environment, but the monitoring coverage that a robot can grasp at a time is limited. In addition, security robots need to be equipped with many sensors so that it deals with all the tasks. Thus, it requires cost. Recently, there is a need for robot systems with lower costs. With this background, I am now applying the parent-child navigation method developed for marine robots to land robots. A robot that can accurately estimate its position acts as a parent robot and supports a child robot with a minimal sensor configuration. Since optical sensors such as cameras are cheaply available these days, robots use cameras to recognize each other’s position and environment. This will become a new monitoring system because it will allow us to deploy many observation systems in an environment.

Recent Research Activities

Figure 1 Deployment scene of 3 AUVs in the sea experiments.

Due to the coronavirus pandemic, research and lectures are being conducted online, but recently we have had more opportunities to conduct them face-to-face while taking infection control measures. To verify the research on parent-child robots, I developed prototype robots with students in the Kuroda laboratory. The robots were developed based on Roomba. Using these robots, we are conducting autonomous navigation tests in the university (Figures 2 and 3), and the students are also conducting their research. For example, there are methods for estimating self-location based on camera images and methods for estimating multiple human trajectories with cameras.

Figure 2 The scene of the experiments in the university.

I am now applying the navigation method I have developed for a new underwater survey by AUVs to land robotics. Although there are differences between underwater and land environments, common methods may solve the problems faced by each of them. I believe that promoting research exchange across fields, without being bound by the boundaries of sea and land, will lead to solving problems that are bottlenecks for both fields.

Figure 4 shows the author’s photo taken together with AUV Tri-TON.

For more details about the activity of Meiji University and Autonomous Mobile Systems Laboratory (Kuroda Laboratory), visit the links shown in the references [2] and [3].

Figure 3 The images obtained by each robot’s camera in the experiments.

References

  • Matsuda, T. Maki, and T. Sakamaki, “Accurate and efficient seafloor observations with multiple autonomous underwater vehicles: theory and experiments in a hydrothermal vent field,” IEEE Robotics and Automation Letters, vol.4, no.3, pp.2333–2339, 2019.
  • Meiji University, https://www.meiji.ac.jp/cip/english/
  • Autonomous Mobile Systems Laboratory

Kuroda Laboratory, https://amslab.tech/