Machine learning used to optimize cyborg cockroach movement in confined spaces

Machine Learning


Have you ever wondered why some insects, like cockroaches, prefer to stay in the dark, or to reduce movement? Some might say it’s a habit of being. Another question is whether it is possible to correct this habit of cockroaches, moving in the dark as they move in bright backgrounds. A scientist at Osaka University may have given a positive answer to this question. They solved this problem by turning cockroaches into cyborgs. They published the study in the journal Cyborg and Bionic Systems.

Over millions of years of evolution, animals in nature have a remarkable ability to survive and thrive in hostile environments. In recent years, these animals have inspired roboticists to develop automated machines that recreate some of these vanished functions: biologically-inspired biomimetic robots. An alternative to this method is to implant stimulation electrodes into the brain or peripheral nervous system to control movement and build controllable machines directly into natural animals called cyborgs. Among these studies, cyborg insects have received increasing attention due to their availability, simpler neuromuscular pathways, and easier manipulation to invasively stimulate the peripheral nervous system or muscles. increase.

Cockroaches have incredible locomotion capabilities, vastly outperforming biomimetic robots of similar size. With such agile locomotion, the cyborg he cockroach is therefore well suited for search and rescue missions in unknown and unstructured environments that are largely inaccessible to conventional robots.

“Cockroaches prefer to stay in dark, confined spaces rather than bright, open spaces. In addition, they tend to be more active in hotter environments.” Mr. Keisuke Morishima, a researcher, explains as follows. For real-time monitoring purposes, it is difficult to install and use a mini live stream he camera in dark or no bright areas. “

“This study aims to optimize the locomotion performance of cyborg cockroaches,” said Morishima. To this end, they proposed a machine learning-based approach to automatically detect the locomotion state of this cyborg cockroach via his IMU measurements. When cockroaches stop or freeze in the dark or cool environment, electrical stimulation is applied to their brains to move them.

“Using this online detector, stimulation is minimal to prevent cockroaches from becoming fatigued from overstimulation.

This idea of ​​limiting electrical stimulation to necessary situations, determined by AI algorithms via on-board measurements, is intuitively promising. “You don’t have to control cyborgs like you control robots. They can have a degree of autonomy that underlies their agile locomotion. You just need to stimulate the cockroaches to turn when you’re standing or to move when you stop unexpectedly,” Morishima said.

“Cyborgs equipped with such a system were able to increase the average search rate and distance traveled by up to 68% and 70%, respectively, while stopping time was reduced by 78%,” said the study authors. . “We have shown that it is possible to apply electrical stimulation to the neck of cockroaches. They can overcome their natural habits, for example in dark and cold environments where they normally move slower. You can increase your movement.”

“In this study, cerci were stimulated to induce free locomotion in Madagascar cockroaches (MHC),” said the study’s authors. In other words, the given stimulus causes the cockroach to move randomly.

Looking to the future, researchers intend to investigate more precise stimulus-motor responses, possibly to control the locomotion direction of cyborg cockroaches, paving the way for real-world search and rescue scenarios.

The research team includes Mohammad Aliyanto, Department of Mechanical Engineering, Graduate School of Engineering, Osaka University, and Chaudhry Mohammad Masum Refat, Department of Mechanical Engineering, Faculty of Engineering, Diponegoro University, Semarang, Indonesia. Kazuyoshi Hirao and Keisuke Morishima, Department of Mechanical Engineering, Graduate School of Engineering, Osaka University.

This research was funded by the Japan Society for the Promotion of Science Grants-in-Aid for Scientific Research (KAKENHI) and the Japan Science and Technology Agency Moonshot Research and Development Program.

DOI 06/02/2023: “Motion Optimization of Cyborg Cockroaches in Bounded Space Incorporating Machine Learning” in the Cyborg and Bionic Systems Journal at https://doi.org/10.34133/cbsystems.0012 has been published.

/Release. This material from the original organization/author may be of a point-in-time nature and has been edited for clarity, style, and length. and do not take a stand. All views, positions and conclusions expressed herein are solely those of the author.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *