Analysis of onboard sensor-based odometry for a quadrotor UAV in outdoor environment

Gabdullin, Aidar; Shvedov, Grigory; Ivanou, Mikhail; Afanasyev, Ilya

doi:10.5954/icarob.2018.gs9-1

Cited by 9 publications

(1 citation statement)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…the measurements’ noise corresponds to the error in distance simulated as

m; the length of the step size in each time-step is simulated as a normal distribution vector with variance of 1m; the odometry noise is taken as

m. Finally, the GNSS observations applied localization with normally distributed noise with zero mean and STD of 30 m; a distorted GNSS receiver applies the localization with a uniform noise of

m that is added to the above normally distributed noise. To illuminate the performance of the proposed scheme, the noises in our simulations were set to be greater than those of common GNSS receivers (e.g., [ 46 ]) and odometry devices (e.g., [ 47 ]). This simulation setup was used for all the experiments throughout the paper.…”

Section: Improved Localization—the Case Of No Disruptionsmentioning

confidence: 99%

Improved GNSS Localization and Byzantine Detection in UAV Swarms

Hacohen

Medina

Grinshpoun

et al. 2020

Sensors

View full text Add to dashboard Cite

Many tasks performed by swarms of unmanned aerial vehicles require localization. In many cases, the sensors that take part in the localization process suffer from inherent measurement errors. This problem is amplified when disruptions are added, either endogenously through Byzantine failures of agents within the swarm, or exogenously by some external source, such as a GNSS jammer. In this paper, we first introduce an improved localization method based on distance observation. Then, we devise schemes for detecting Byzantine agents, in scenarios of endogenous disruptions, and for detecting a disrupted area, in case the source of the problem is exogenous. Finally, we apply pool testing techniques to reduce the communication traffic and the computation time of our schemes. The optimal pool size should be chosen carefully, as very small or very large pools may impair the ability to identify the source/s of disruption. A set of simulated experiments demonstrates the effectiveness of our proposed methods, which enable reliable error estimation even amid disruptions. This work is the first, to the best of our knowledge, that embeds identification of endogenous and exogenous disruptions into the localization process.

show abstract

“…the measurements’ noise corresponds to the error in distance simulated as

m; the length of the step size in each time-step is simulated as a normal distribution vector with variance of 1m; the odometry noise is taken as

m. Finally, the GNSS observations applied localization with normally distributed noise with zero mean and STD of 30 m; a distorted GNSS receiver applies the localization with a uniform noise of

Section: Improved Localization—the Case Of No Disruptionsmentioning

confidence: 99%