Scu| gait switching and target navigation of micro swimming robot through deep reinforcement learning

【 title 】Gait switching and targeted navigation of microswimmers via deep reinforcement learning

【 The author team 】Zonghao Zou, Yuexin Liu, Y.-N. Young, On Shun Pak, Alan C. H. Tsang

【 Date of publication 】2022.6.21

【 Thesis link 】https://www.nature.com/articles/s42005-022-00935-x.pdf

【 Recommended reasons 】 The micro swimming robot switches between moving gait , Implement complex navigation strategies , Such as running and tumbling , To explore their environment and search for specific targets . This ability to navigate targets through adaptive gait transitions is important for developing the ability to perform complex biomedical tasks autonomously ( Such as targeted drug delivery and microsurgery ) The intelligent artificial micro swimming robot is particularly ideal . This paper uses the method of deep reinforcement learning , The model micro swimming robot can self learn effective gait , Including translation 、 Rotation and combined motion . Artificial intelligence (AI) The driven swimming robot can adaptively switch between various gait , To navigate to the target location . The multi-mode navigation strategy reminds people of the gait switching behavior adopted by the micro swimming robot . The results show that , these AI Driven micro swimming robot in unpredictable 、 It has great application potential in complex fluid environment .