Fault Tolerant Attitude Estimation Strategy for a Quadrotor UAV under Total Sensor Failure

Abstract

In this paper, we propose a conventional extended Kalman filter (CEKF) based new fault-tolerant scheme for the control of quadrotor drones suffering from a total failure of gyroscope sensor. The fault-tolerant conventional extended Kalman filter (FTCEKF) combines two conventional extended Kalman filters under the name CEKF1 and CEKF2 as well as a fault detection algorithm. In the case of nominal operation of the quadrotor, it is possible to obtain a sufficiently good estimation of the desired attitude, position and speed using the first conventional extended Kalman filter (CEKF1). However, in the presence of a total failure of the gyroscope, this filter leads to the collapse of the whole system. To overcome this type of scenarios, a second filter is used as a backup filter (CEKF2), where its role is to ensure a satisfactory attitude estimation using only a low cost magnetometer. To detect a gyroscope sensor failure, a reconfiguration mechanism based on a probabilistic algorithm to make a decision to switch to the backup filter is proposed in this article. The results show that the proposed (FTCEKF) scheme is approved to be robust against a total gyroscope failure with higher filtering performance.