目录
AHRS Attitude Heading Reference System
RTK-enabled GNSS Real Time Kinematics
手动陀螺偏差估计/无旋转更新 Manual Gyro Bias Estimation(MGBE) / No Rotation Update
4.3 本地磁偏Local magnetic declination
1 Xsens传感器的海洋应用Marine Subsea:
- ROV (Remotely Operated Vehicle)
- AUV (Autonomous Underwater Vehicle)
- BUOY航标
- Transponders应答器(跟踪海洋生物用的)
- Seafloor Mapping海床制图
2 传感器简述
IMU Inertial Measurement Unit
VRU Vertical Reference Unit
AHRS Attitude Heading Reference System
GNSS GPS-enabled Initial Navigation System
RTK-enabled GNSS Real Time Kinematics
RTK-enabled VINS Visual Initial Navigation System
海洋水下环境,不能使用GNSS,只有IMU/VRU/AHRS
水下环境的挑战
航偏角yaw错误/漂移
成本
4 解决方案
4.1 磁干扰Magnetic interference
AUV本体产生的磁场(homogeneous magnetic field)如下示意图,该磁场稳定不变,可以通过标定补偿。
下面以Xsens的磁干扰标定为例。
当进行磁性校准时,在将传感器安装到探测器内部后,尽可能在三维空间中移动AUV,以考虑到三维空间内尽可能多的点,从而提供更多校准数据。这些数据被反映在Xsens界面的球形图中,球面上的点即为标定中得到的数据点。某些情况下三维标定不可行,如重型AGV只能在地面移动,这种情况仅进行二维标定也是可以的。但对于水下应用的AUV则可以进行3D磁标定。
界面右侧是标定结果。黑色线校准前,蓝色的线是校准后的结果。我们在这里关注的是1范数。标定后的范数稳定在1附近(即稳定在圆球球面上)
4.2 陀螺仪偏差Gyro bias
手动陀螺偏差估计/无旋转更新 Manual Gyro Bias Estimation(MGBE) / No Rotation Update
这个功能的基本作用是通过UHRS设备向IMU传达信息。当传感器静止,偏航yaw应该是稳定的。该命令会告知传感器在接下来的5秒内,传感器将保持静止,偏航角保持不变。
由于机械应力/振动引起的陀螺仪bias:
- 橡胶阻尼器。吸收大部分的振动和冲击,避免这些振动和冲击直接传递到传感器上。
- 对于表面贴装设备SMDs,例如 Xsens MTI-1,使用回流焊接工艺,而不是手工焊接或PLCC插座。手工焊接或PLCC插座可能会引入机械应力,增加角速度bias。
4.3 本地磁偏Local magnetic declination
方法:将最新的世界磁场模型 (World Magnetic Model, WM) 存储在板载存储空间,手动存储当前设备位置(经纬度/海拔),以便该设备可以准确地参考其相对于当地磁场的偏航。
4.4 成本
- 光纤陀螺仪 (Fiber Optic Gyroscopes,FOG) 和环形激光陀螺仪 (Ring Laser Gyroscopes, RLG) 体积庞大且昂贵。
- 声呐同样维护成本高
- 声定位系统Acoustic positioning system,如LBL,SBL,需要部署转发器
- 基于MEMS的IMU、VRU、AHRS:便宜,小,支持多种通信
水下探测器例子
Carnegie Mellon University (CMU) 团队开发了TartanAUV,这是一种自主潜艇,使用MTi VRU 进行精确的水下导航。
挑战
"Drift in the dead reckoning calculations" 指的是在使用航位推算(Dead Reckoning, DR)方法进行导航时,由于各种因素导致的累积误差或偏移。航位推算是一种基于已知起点、航向、航速和时间来估算当前位置的方法,不依赖于外部信号(如GPS)。然而,由于实际航行中可能存在的各种不确定性,如风向、水流、船体姿态变化、速度测量误差等,这些因素会逐渐累积并导致计算出的位置与实际位置之间产生偏差,这种偏差就被称为“漂移”。
简单来说,就是在没有外部参照物校正的情况下,仅凭航向和速度来估算位置,时间一长,这种估算就会因为各种因素而变得不那么准确,产生的偏差就是“Drift in the dead reckoning calculations”。
方案
- MTi VRU 准确姿态和航向数据
- Xsens ROS1/ROS2 驱动与 Nvidia Jetson 平台兼容
- Xsens 专有传感器融合算法和同步,鲁棒性好
- VRU测量(最高400Hz)填补了低频DVL数据之间的空白
6 提问
Q: 世界磁场模型是如何存储在你的设备上的? How is the world magnetic model stored on your devices?
A: 世界磁场模型嵌入在固件中。因此每当有固件更新时,固件都会包含最新的世界磁场模型。So the world magnetic model is embedded inside the firmware itself. So we make sure that we whenever there's a firmware update, we make sure that the firmwares has the latest world magnetic model.
Not only in the sea but outside also. Do we need to calibrate the IMU compulsory?
A:是的,特别是如果您使用的是提供参考的AHRS设备,那么校准它非常重要。进行磁场校准,以便您能够考虑均匀的磁场。这将考虑您车辆内的所有磁场,例如您正在使用的探测器。因此,强烈建议进行磁场校准。进行磁校准有多种方法。您可以使用工具,也可以使用SDK中的命令。
Yeah, if, especially if you're using an AHRS device which uses which provides a reference, it is very important to calibrate it. Perform a magnetic calibration so that you can account for the homogeneous magnetic fields. This will take into account all the magnetic fields within your vehicle, for example a Rover that you're using. So yeah, it is highly recommended to perform a magnetic calibration. And there are multiple ways to perform a magnetic calibration too. You can use the tool or you can also use commands from the SDK as well.
Q:您主张使用惯性测量单元(IMU)而不是声学参考系统。您如何在没有绝对参考的情况下解决漂移补偿问题?
You argue for using IMU over acoustic reference systems. How do you solve drift compensation without an absolute reference?
A:这绝对不是单一的选择。如果您想要一个低成本的解决方案,既低成本又准确的水下导航稳定解决方案,您可以使用基于MEMS的技术。基本上,对于某些需要非常高精度的应用,并且对相关成本完全可以接受,例如声学参考系统,在这些应用中,您确实可以在惯性测量单元(IMU)之外使用声学参考系统。然而,寻找低成本解决方案的客户可以选择我使用的主要空间。
So it's definitely not one or the other. If you want a low cost solution, low cost and accurate solution for underwater navigation stabilization, you can use MEMS based. Basically, however, for certain applications which require very high accuracy and also are completely fine with the costs that are associated with, for example, acoustic reference systems, in those kind of applications you can certainly use the acoustic reference systems in addition to the IMUs. However, people customers looking for the low cost solutions can go for just the main space I use.
Q: 是否可以通过ROS包将经纬度更新到设备上?
Can the lat long be updated to the device through the ROS package?
A: 是的,您可以通过该包更新传感器的位置,基本上就是更新经纬度。您可以使用ROS包,也可以使用我们的SDK命令。
Yes, you can update the location of the sensor, basically the lat long using the package itself.
The Ross package or you can use our SDK commands as well.
Q: 能否将校准存储在非易失性存储器中并进一步使用吗?
can we store the calibration in NV memory and use further?
A: 一旦您执行了磁性校准,就有一个选项可以将新的校准数据写入设备的内存中。之后,它将计算新的磁校准值。您还可以在我们空的管理应用程序中的建模参数内查看新的校准值。
So once you perform a magnetic calibration, there's an option that you can write the calibration the new calibration data into the memory itself, memory of the device and. After which it will account for the new magnetic calibration values. You can also see the new calibration values inside the modeling parameters in our empty Manager application.
Q: 关于世界磁场模型,你们在设备中实现了世界磁场模型的方程,并且每五到七年进行一次软件更新?
about world magnetic model. You implement world magnetic model equations in your device and updated every five or seven years with some software update too.
A: 是的,基本上我们在固件中包含或实现了世界磁场模型。每当世界磁场模型有更新时,它也会包含在固件更新中。一旦您安装了新的固件,如果全球磁模型有更新,它将包含该更新。
Yeah. So basically we include or implement the world magnetic model within our firm within the firmware itself. And yes, whenever there is an update in the world magnetic model, it is it is included within the firmware update itself as well. So once you install the new firmware, if there is an update in the world magnetic model it will include that.
Q:果我们看到磁校准变化很快,有什么可用的过程可以遵循,以避免磁力计精度的不准确?
if we see the magnetic calibration changes rapidly, what is the process available to follow to avoid the inaccuracy of magnetometer accuracy?
A: 磁校准主要或应该主要用于考虑来自最终应用单元本身的磁校准,例如测量仪。如果你所处的环境存在大量外部磁场或不均匀磁场。这些磁场无法通过磁校准来解释。这种情况最好使用VRU,它在不使用磁力计进行读数的情况下提供无参考值。偏航角在您启动设备时从0开始。
So the magnetic calibration is mainly or should mainly be used with the to account for the magnetic calibration from the end application unit itself, like the Rover. If your environment within which you're operating in.
Q: 在像AUV应用这样的狭小空间工作并不容易。传感器由于电干扰无法使用。有解决方案吗?
It turns out that working in tight spaces as in AUV applications heading is not. Unusable due to electric interference. Is there a solution to overcome?
A: 是的,例如高压电缆会造成电磁干扰。因此,我们通常建议安装MTI设备。至少要远离高压电缆一点,比如至少10到15厘米,并且在此基础上进行磁场校准。
Yeah, with electrical interference, for example, high power cables. So we usually recommend to mount the MTI devices. Of at least a little bit away from the high power cables, for example about 10 to 15 centimeters at least, and also on top of that perform magnetic calibration.
Q: 在将AHRS信息与其他传感器(如DVL)融合用于水下应用时,推荐的融合方法或滤波器结构是什么?例如扩展卡尔曼滤波器(EKF)或无迹卡尔曼滤波器(UKF)或其他方法?
If fusing the AHRS info with other sensors like a DVL for subsea application, what is the recommended fusion method or filter structure? Example Ek extended Kalman filter or UK or something else?
A: 我们有自己的专有传感器融合算法,这种算法只能在我们的设备中工作,并用于同步。其他传感器,如DVL、摄像头等。我们提供同步功能。所以有多种同步功能可以使用。我有一篇文章。如果您访问我们的知识库并输入与MTI设备的同步相关的内容,您将找到一篇文章,讨论MTI设备提供的所有同步功能,以及如何将其他设备与MTI进行同步的示例。
So we have our proprietary sensor fusion algorithm and that is only that can only work within our devices and for synchronization with. Other sensors such as DVL, cameras, etc. We offer synchronization functionality. So and there are multiple synchronization functions that you can use. I have an article. Which if you go to our knowledge base and type synchronization with MTI devices, you can find an article which talks about all the synchronization functions that the MTI devices offer and also examples of how a synchronization how you can synchronize other devices with MTI.