Stable Heteroclinic Channels (SHCs) are a novel framework for creating central pattern generators (CPGs). In MATLAB 2022a, we implement an SHC controller for lateral undulation in a simulated aquatic serpentine robot with six segments. We demonstrate that anguilliform locomotion, wherein lateral displacement during undulation increases from head to tail, can also be achieved by scaling SHC parameters. When compared with a sine controller, SHCs achieve equivalent velocity and 42% lower overall COT for the same lateral undulation gait. COT improvements are attributable to lower startup energy consumption. We propose a method for integrating tactile sensing into the SHC trajectory planner by scaling the kernel activation rate. We show that the sensorized SHC controller exhibits faster, lower amplitude gait cycles, enabling progression in a narrow crevice. Compared with a sensorized sine controller, the sensorized SHC controller achieves 41% greater speed, but 52% lower efficiency. This research demonstrates that SHCs benefit from intuitive tuning like sine control, while also producing smooth trajectories representative of other CPG frameworks.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Simulated Control of an Aquatic Serpentine Robot with Stable Heteroclinic Channels

  • Nathaniel Mengers,
  • Natasha Rouse,
  • Kathryn A. Daltorio

摘要

Stable Heteroclinic Channels (SHCs) are a novel framework for creating central pattern generators (CPGs). In MATLAB 2022a, we implement an SHC controller for lateral undulation in a simulated aquatic serpentine robot with six segments. We demonstrate that anguilliform locomotion, wherein lateral displacement during undulation increases from head to tail, can also be achieved by scaling SHC parameters. When compared with a sine controller, SHCs achieve equivalent velocity and 42% lower overall COT for the same lateral undulation gait. COT improvements are attributable to lower startup energy consumption. We propose a method for integrating tactile sensing into the SHC trajectory planner by scaling the kernel activation rate. We show that the sensorized SHC controller exhibits faster, lower amplitude gait cycles, enabling progression in a narrow crevice. Compared with a sensorized sine controller, the sensorized SHC controller achieves 41% greater speed, but 52% lower efficiency. This research demonstrates that SHCs benefit from intuitive tuning like sine control, while also producing smooth trajectories representative of other CPG frameworks.