This paper deals with a positioning method of a vehicle on undulating ground using GPS and internal sensors (a fiber optic gyro, a roll pitch sensor and wheel encoders). To make the best use of the sensory data, the following five factors must be considered.
(1) The effect of undulation to yaw sensing.
(2) The drift of the gyro.
(3) The delay and scarcity of the GPS data.
(4) The inconstancy of travel distance per encoder pulse due to the differential mechanism and the slip of the wheels.
(5) The effect of the position of the GPS antenna on position estimation.
We propose an observation model of the GPS and the internal sensors, which considers these five factors. A positioning method was developed on the basis of the observation model. A prototype of autonomous mower was produced to evaluate this positioning method. The relationship between the GPS accuracy and the sensor fusion system accuracy is discussed. The experimental results show that the sensor fusion system was more accurate than GPS only or internal sensor system only. If the accuracy of the GPS was 1m, the accuracy of the sensor fusion system was 0.2m.
抄録全体を表示