基于仿生侧线传感器的机器鱼游速测量方法研究

The bio-inspired robot fish has the potential advantage of low power consumption, low noise, and good maneuverability, which were paid very much attention. However, the velocity, which is a basic parameter of the manipulation and control, is hard to measure. This project was inspired by the fish's lateral line apparatus which is used to detect the liquid changing around, and is going to design a kind of flexible sensor to measure the relative velocity of liquid on the robot fish's surface. Then, we plane to fuse the information from the sensor array to calculate the state of the liquid and the velocity of robot fish. The research work divides into two aspects. One is the sensor design, including the bio-inspired design, the velocity measurement between solid and liquid in low speed and open flow field, the yawing noise infection on the measurement. The other aspect is the study on the sensor array based on the fish's lateral line distributing and the character of robot fish, including the fuse method to calculate the velocity from the multiple data, and fuse the data with heterogeneous pose sensor data. At last, we will build a simulating platform and improve robot fish and vision based measure platform in existence to verify the validity of our method. The research works of this project can develop the theory support and the technique foundation for the bio-inspired lateral line for the underwater robot.

仿生机器鱼具有低功耗、低噪音、机动灵活等潜在优点而受到广泛关注，但是作为运动性能和控制的重要参数，运动速度，这一基本参数的测量却有很大的困难。本课题组受到鱼类的感知环境流体变化的器官――侧线的启发，采用柔性弯曲传感器在机器鱼表面特定点测量流体的相对速度，并通过融合多个传感器的阵列信息推算机器鱼周边的流体状态，进而计算机器鱼的速度。研究内容一方面是传感器的仿生设计，研究其在低流速、开放流场中的流体与固体分界面上的相对速度测量，并计算机器鱼在仿生波动推进条件下，艏摇噪声对测量的影响；另一方面要结合生物侧线分布特征和机器人的特性设计传感器阵列，研究从多组数据中推算实际速度值的信息融合方法，并和其他位姿传感器进行异质信息数据融合。建立仿真平台并改进现有机器鱼和视觉测量平台，验证以上方法的有效性，为仿生侧线传感器在水下机器人领域的广泛应用提供必要的理论支持和技术基础。

本项目正对仿生机器鱼运动速度，这一基本参数的测量的困难，研究了一种仿生侧线系统。受到鱼类的感知环境流体变化的器官――侧线的启发，采用柔性弯曲传感器在机器鱼表面特定点测量流体的相对速度，并通过融合多个传感器的阵列信息推算机器鱼周边的流体状态，进而计算机器鱼的速度。一方面完成了传感器的仿生设计，研究其在低流速、开放流场中的流体与固体分界面上的相对速度测量，并计算机器鱼在仿生波动推进条件下，艏摇噪声对测量的影响；另一方面结合生物侧线分布特征和机器人的特性设计传感器阵列，研究了从多组数据中推算实际速度值的信息融合方法，并和其他位姿传感器进行异质信息数据融合。建立仿真平台并改进现有机器鱼和视觉测量平台，验证以上方法的有效性，为仿生侧线传感器在水下机器人领域的广泛应用提供必要的理论支持和技术基础。