Bionic FlowSight Sensor Enables Underwater Robots to Perceive Water Currents

Inspired by the lateral line system found in fish, which allows them to navigate and avoid obstacles in turbulent underwater environments, researchers at the Institute of Automation, Chinese Academy of Sciences, have developed a novel bio-inspired sensor called FlowSight. This innovative sensor provides water robots with precise "hydrodynamic perception" capabilities, opening new avenues for autonomous navigation and environmental monitoring in complex aquatic conditions. FlowSight mimics the neuromasts, or sensory structures, found along the sides of fish. The sensor consists of a flexible silicon whisker that detects changes in water flow. When water currents impact this whisker, its deformation is captured as a sequence of high-resolution images by an integrated camera. A deep learning model then processes these images to extract data on water velocity and direction, achieving accurate single-point flow vector sensing without additional equipment. The relative error for measuring water speed is just 3.05%, and the error for measuring direction is 0.98%. The research team successfully integrated FlowSight into a biomimetic underwater robot named RoboDact. Through a series of closed-loop motion control experiments, they demonstrated that RoboDact can navigate against currents and dynamically adjust its posture, much like a real fish. This capability significantly enhances the robot's adaptability to various underwater environments, making it an invaluable tool for underwater exploration and ecological monitoring. The detailed findings of this research were published in the prestigious journal IEEE Transactions on Robotics under the title "FlowSight: Vision-Based Artificial Lateral Line Sensor for Water Flow Perception." The work has been supported by major national science and technology projects, including the National Science and Technology Major Project on Brain Science and Brain-inspired Research, the National Natural Science Foundation of China, and the China Postdoctoral Science Foundation. In practical tests, FlowSight sensors performed admirably in variable-speed swim lanes, accurately estimating both water speed and direction. The integration of FlowSight into RoboDact enabled the robot to adjust its speed and orientation in real-time based on the detected water flow, thus ensuring its ability to operate effectively in diverse and challenging underwater settings. This breakthrough in bio-inspired sensors marks a significant step forward in the field of underwater robotics, combining the precision of visual sensing with the robustness of deep learning models. FlowSight's unique design and performance highlight its potential to revolutionize how robots perceive and interact with their aquatic surroundings, paving the way for more advanced and versatile underwater technologies. For further details, the paper authored by Tiandong Zhang, Rui Wang, Qiyuan Cao, Shaowei Cui, Gang Zheng, and Shuo Wang, titled "FlowSight: Vision-Based Artificial Lateral Line Sensor for Water Flow Perception," is available in the IEEE Transactions on Robotics in 2025.