Liu, YufeiNogichi, NoboruIshii, Kazunobu2021-09-242021-09-242018-03-31https://doi.org/10.15832/ankutbd.446369http://hdl.handle.net/20.500.12575/74905The objective of this study was to develop a small-sized and low-cost unit to provide attitude measurements for lightloaded, small-sized and cost effective agricultural robot application. The attitude measurement unit comprised an electronic control unit (ECU) and a gyroscope and an accelerometer within a small-sized and low-cost IMU. In order to avoid the measurement limitations of a single sensor, a self-adaptive complementary filter and a Kalman filter were discussed and compared for sensor fusion. By comparison, in respect of preventing angle drift and maintaining dynamic characteristics, the Kalman filter has the significant advantage, especially in dynamic motion. In the comparison with a highly precise aviation-level fiber optic gyroscope (FOG), the results showed that the static angle drift was restrained by Kalman filter which reached the performance of the FOG. And in the series of farm experiments, the dynamic characteristic of the developed attitude measurement unit is close to the FOG performance in the sub-degree level. This is an acceptable accuracy for light-loaded, small-sized and cost effective agricultural robot application such as agriculture drone, greenhouse robots, harvesting robot arm and so on.enIMU sensorComplementary filterKalman filterArticle24133412148-9297