State Estimate Recovery for Autonomous Quadcopters

Abstract

A method for recovery from the complete loss of the state estimate is presented for autonomous quadcopters. Given an aerodynamic force model, the only measurements used to reinitialize the state estimate by means of a bank of extended Kalman filters are the angular rate and linear acceleration measurements of an IMU. The method is integrated within a complete recovery logic on a quadcopter platform and experimentally evaluated. I. INTRODUCTION Over the course of the last two decades, unmanned aerial vehicles (UAVs), and particularly quadcopters, have received increasing attention. Low-cost, lightweight sensing technology paired with powerful processors enabled their widespread use as an experimental platform for research in the field of robotics, as well as in industry: for example for aerial delivery, inspection, surveillance, photography, mapping, inventory management and entertainment. With the emergent use of quadcopters, their safety and robustness have become important requirements. Ideally, a quadcopter is completely fault-tolerant, meaning that it is engineered in such a way that a failure affecting any of the components of the system does not prevent the intended operation being completed. An intuitive concept to create a fault-tolerant system is redundancy. By duplicating every component, a quadcopter can continue operation relying on the backup components, which almost completely eliminates the risk of a catastrophic failure. The major drawbacks of redundancy, however, are the increased cost, complexity and weight of the system. Redundancy for actuator fault-tolerance can alternatively be achieved by duplicating only a subset of components: For example, in [1], [2], [3] and [4] fault-tolerant control for hexacopters is studied. Another alternative way to achieve actuator redundancy is to introduce different means of actuation, for example by rotors that can be actively tilted [5]. In the context of sensor fault-tolerance, redundancy is typically not achieved by duplication of components, but by a sensor configuration that provides redundant information, see for example [6]. A summary of work on sensor and actuator fault-tolerance for quadcopters can be found in [7]. A less strict requirement is a fail-safe design: Such a system is designed to respond to a failure by taking an action that minimizes damage and/or harm to the equipment and its environment. For example, in [8] a control strategy is proposed that allows a quadcopter to continue controlled flight despite the loss of up to three propellers, without