Position and attitude estimation using ultrasonic waves for autonomous flying robots and system construction

It is important for an autonomous flying robot to be able to estimate self-position and attitude. However, GPS cannot be used because of the accuracy problem indoors. This paper proposes a measurement system using the positional relation between an ultrasonic transmitter placed on the ground and an ultrasonic receiver mounted on a flying robot. Moreover, the system applies an estimator using an Extended Kalman Filter (EKF) in order to improve the measurement accuracy. The state variable consists of the position vector and attitude vectors. In applying the EKF, a system model is formulated theoretically assuming that the measured distance data are random variables with uniform distribution. Theoretical derivation reveals that the variance calculation becomes very complicated. The simulation experiment demonstrates the propriety of the theory. In addition, the design of a prototype board and the implementation problem with the hardware are discussed.