LiBeamsNet AUV Velocity Vector Estimation in Situations of Limited DVL Beam Measurements Nadav Cohenand Itzik Klein

2025-05-03 0 0 1.5MB 5 页 10玖币

侵权投诉

LiBeamsNet: AUV Velocity Vector Estimation in

Situations of Limited DVL Beam Measurements

Nadav Cohen∗and Itzik Klein

The Hatter Department of Marine Technologies

Charney School of Marine Sciences, University of Haifa

Haifa, Israel

Abstract—Autonomous underwater vehicles (AUVs) are em-

ployed for marine applications and can operate in deep under-

water environments beyond human reach. A standard solution

for the autonomous navigation problem can be obtained by

fusing the inertial navigation system and the Doppler velocity

log sensor (DVL). The latter measures four beam velocities to

estimate the vehicle’s velocity vector. In real-world scenarios,

the DVL may receive less than three beam velocities if the

AUV operates in complex underwater environments. In such

conditions, the vehicle’s velocity vector could not be estimated

leading to a navigation solution drift and in some situations

the AUV is required to abort the mission and return to the

surface. To circumvent such a situation, in this paper we propose

a deep learning framework, LiBeamsNet, that utilizes the inertial

data and the partial beam velocities to regress the missing

beams in two missing beams scenarios. Once all the beams are

obtained, the vehicle’s velocity vector can be estimated. The

approach performance was validated by sea experiments in the

Mediterranean Sea. The results show up to 7.2 % speed error

in the vehicle’s velocity vector estimation in a scenario that

otherwise could not provide an estimate.

Index Terms—Autonomous underwater vehicle (AUV), Inertial

navigation system (INS), Doppler velocity log (DVL), Deep

Learning

I. INTRODUCTION

The primary purpose of the autonomous underwater vehicle

(AUV) is to perform scientiﬁc tasks in the great depths of the

ocean. In order to do so, the AUV has to operate autonomously

in the underwater environment. Thus, it contains several sen-

sors collecting data to enable autonomous navigation [1].

The inertial navigation system (INS) provides the navigation

solution utilizing inertial measurements from its three-axis

accelerometer and a three-axis gyroscope; providing the spe-

ciﬁc force and angular velocity vectors, respectively [2], [3].

The navigation solution gives the AUV’s position, velocity,

and orientation. Although this information is sufﬁcient for au-

tonomous navigation, the INS cannot be used as a standalone

solution due to its nature to accumulate error over time [4],

[5]. To that end, a Doppler velocity log (DVL) is typically

used in the AUV to overcome this problem.

The DVL is based on the Doppler effect, while the bottom-

lock refers to a situation where four beams are transacted to

the seaﬂoor and reﬂected back to the sensor. The DVL can

achieve a typical velocity measurement accuracy of 0.2% of

∗Corresponding author: N. Cohen (email: ncohe140@campus.haifa.ac.il).

the current velocity and is, therefore, considered an accurate

sensor [6]. In most situations, an accurate navigation solution

can be obtained by performing a fusion between the INS and

the DVL for error accumulation avoidance [7]–[9]. However,

if the process noise covariance of the ﬁlter is not determined

properly, the navigation solution may drift. To circumvent

such situations both model and learning adaptive ﬁlter solution

exists in the literature [10], [11]. As the DVL is the aiding

sensor to the navigation solution, it is crucial to obtain con-

tinuous DVL data ﬂow. However, in real-world scenarios, the

DVL may receive only partial beam measurements in different

situations. These situations include sea creatures blocking the

acoustics beams, trenches on the seaﬂoor, and extreme pitch

and roll maneuvers (such as diving) [12], [13]. The minimum

amount of measured beams for obtaining an AUV velocity

estimate is three. If there are fewer, an AUV velocity vector

cannot be derived, which results in an error accumulation of

the navigation solution. Usually, in this situation, the AUV

will abort the mission and will be forced to surface.

To cope with such situations, several solutions were suggested

in the literature. In [14] additional data and the current partial

beams measurements were employed to create virtual beams

enabling the estimation of the AUV velocity. In [15], the

partial beams were used directly in the INS/DVL sensor fusion

in a tightly coupled (TC) approach. Furthermore, a data-driven

method was used to compensate for a situation of one missing

beam using past DVL measurements and Long short-term

memory (LSTM) model and outperformed the model-based

approach [16]. Recently, the INS/DVL performance using a

tightly coupled approach in situations of partial DVL beam

measurement, was improved using a learning virtual beam

aided solution [17].

In situations of complete DVL measurement, we proposed

a deep learning approach, named BeamsNet, to estimate the

AUV velocity vector and replace the model-based approach

[18]. We demonstrated our approach’s ability to outperform

the model-based approach using a dataset collected from

sea experiments made with an AUV. Leveraging from that

work, in this paper, a deep learning approach, LiBeamsNet

(LImited Beams NETwork), is proposed to compensate for

a situation of two missing DVL beams. Our network uses

inertial readings and the current partial beams as input and

provides the two missing beams. Once all the beams are

arXiv:2210.11572v1 [cs.RO] 20 Oct 2022

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

LiBeamsNet:AUVVelocityVectorEstimationinSituationsofLimitedDVLBeamMeasurementsNadavCohenandItzikKleinTheHatterDepartmentofMarineTechnologiesCharneySchoolofMarineSciences,UniversityofHaifaHaifa,IsraelAbstractAutonomousunderwatervehicles(AUVs)areem-ployedformarineapplicationsandcanoperateindeepunder...

展开>> 收起<<

LiBeamsNet AUV Velocity Vector Estimation in Situations of Limited DVL Beam Measurements Nadav Cohenand Itzik Klein.pdf

共5页,预览1页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

LiBeamsNet AUV Velocity Vector Estimation in Situations of Limited DVL Beam Measurements Nadav Cohenand Itzik Klein

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: