Alternative approaches for UAV dead reckoning based on the immunity paradigm

• Published in: Aerospace science and technology Vol. 98; p. 105742
• Main Authors: Alnuaimi, Mohanad, Perhinschi, Mario G.
• Format: Journal Article
• Language: English
• Published: Elsevier Masson SAS 01.03.2020
• Subjects: Artificial immune system; GNSS-denied environment; UAV autonomous flight; UAV autonomous flight; GNSS-denied environment; Artificial immune system
• ISSN: 1270-9638; 1626-3219
• DOI: 10.1016/j.ast.2020.105742

The immunity paradigm has been recently investigated as a potential solution to the problem of unmanned aerial vehicles (UAV) navigation when external sensor systems and information of opportunity are not available. The effectiveness of specialized immunity cells in memorizing the characteristics of invading antigens, as well as the characteristics of the response of the host organism, or the antibodies, represent the source of inspiration for the development of an on-board intelligent system, an artificial immune system (AIS). The AIS is expected to provide corrections to a simple dead reckoning algorithm, such that autonomous trajectory tracking control of a UAV can be maintained over extended periods of time. Variables that can characterize the dynamic state of the vehicle and can be measured exclusively with on-board sensors are assimilated to the antigens, while the antibodies consist of the adjustments applied to the integration algorithm, such that accurate estimations of vehicle position and velocity are obtained. The AIS is built as a collection of instantaneous values of antigens and corresponding antibodies, referred to as artificial memory cells. The AIS is constructed using data collected under nominal conditions, when all sensors and systems function correctly and vehicle position and velocity are available. Two alternative approaches for building and operating the AIS are implemented and investigated in this paper. One relies on providing corrections to the output and the other to the input of the integration scheme. The AIS for both scenarios have been generated and tested using the West Virginia University unmanned aerial systems simulation environment. Both alternative approaches have been demonstrated to achieve similar performance in terms of significantly improving the vehicle position and velocity estimation algorithm and leading to desirable trajectory tracking and mission completion.