Abstract This paper presents an optimum structural design method for a sensor capable of measuring the control torque of the micro flapping-wing robot. Torque measurement in flapping-wing robots has special demands on the sensor bandwidth. The method in this paper focuses on the development of a multi-objective optimization method based on the surrogate model to solve the sensor indicator design issues, for example, increasing sensitivity while meeting bandwidth requirements. Initially, Latin hypercube sampling was applied to the choice of the characteristic parameters of the sensor. Based on finite element techniques, the surrogate models describing displacement and eigenfrequency were established to characterize the sensor indicators, and a multi-objective optimization was conducted. According to the optimization results, the sensor structure was manufactured, and a torque measurement platform was set up. Calibration experiments and frequency response experiments demonstrated a high level of consistency between the surrogate model and the experimental data. With the assistance of the eddy current displacement sensor, the actual displacement sensitivity coefficient of the torque sensor is determined to be 0.4325 mm mNm−1, consistent with the calculation of the surrogate model. The characteristic frequency is 488.5 Hz, with a relative error of 7.47% compared to the surrogate model. This indicates that the optimum structural design method can be utilized for the rapid design of torque sensors for special requirements, thus laying the foundation for the control torque measurement of micro flapping-wing robots.
Read full abstract