A correct calibration of the sensor components inside the MTi is essential for an accurate output. The quality and importance of the calibration are of the highest priority and so each Xsens MTi is calibrated and tested by subjecting each product to a wide range of motions and temperatures.
The individual calibration parameters are used to convert the sensor component readout (digitized voltages) to physical quantities as accurately as possible, compensating for a wide range of deterministic errors. Additionally, the calibration values are used in Xsens sensor fusion algorithms, as discussed below.
The orientation of the VRU/AHT and AHRS is computed by Xsens Kalman Filter. XKF3™ is a proven sensor fusion algorithm, which can be found in various products from Xsens and partner products. The industrial applications version is XKF3i: it uses signals of the rate gyroscopes, accelerometers and magnetometers to compute a statistical optimal 3D orientation estimate of high accuracy with no drift for both static and dynamic movements.
The design of the XKF3i algorithm can be summarized as a sensor fusion algorithm where the measurement of gravity (by the 3D accelerometers) and Earth magnetic north (by the 3D magnetometers) compensate for otherwise slow, but unlimited, increasing (drift) errors from the integration of rate of turn data (angular velocity from the rate gyros). This type of drift compensation is often called attitude and heading referencing and such a system is referred to as an Attitude and Heading Reference System (AHRS).
XKF3i stabilizes the inclination (i.e. roll and pitch combined) using the accelerometer signals. An accelerometer measures gravitational acceleration plus acceleration due to the movement of the object with respect to its surroundings.
XKF3i uses the assumption that, on average, the acceleration due to the movement is zero. Using this assumption, the direction of the gravity can be observed and used to stabilize the attitude. The orientation of the MTi in the gravity field is accounted for so that centripetal accelerations or asymmetrical movements cannot cause a degraded orientation estimate performance. The key here is the amount of time over which the acceleration must be averaged for the assumption to hold. During this time, the rate gyroscopes must be able to track the orientation to a high degree of accuracy. In practice, this limits the amount of time over which the assumption holds true.
However, for some applications, this assumption does not hold. For example, an accelerating automobile may generate significant accelerations for time periods lasting longer than the maximum duration during which the MT’s rate gyroscopes can reliably keep track of the orientation. This will degrade the accuracy of the orientation estimates with XKF3i somewhat, because the application does not match the assumptions made in the algorithm. Note, however, that as soon as the movement again matches the assumptions made, XKF3i will recover and stabilize. The recovery to optimal accuracy can take some time.
NOTE: To be able to accurately measure orientations as well as position in applications which can encounter long-term accelerations, we offer a solution that incorporates a GNSS receiver, the MTi-G-710 GNSS/INS.
By default, yaw is stabilized using the local (earth) magnetic field (only in the AHRS product types). In other words, the measured magnetic field is used as a compass. If the local Earth magnetic field is temporarily disturbed, XKF3i will track this disturbance instead of incorrectly assuming there is no disturbance. However, in the case of structural magnetic disturbance (>10 to 30 s, depending on the filter setting) the computed heading will slowly converge to a solution using the 'new' local magnetic north. Note that the magnetic field has no direct effect on the inclination estimate.
In the special case the MTi is rigidly strapped to an object containing ferromagnetic materials, structural magnetic disturbances will be present. In that case, there are solutions to use the magnetometers after all. Please refer to BASE: Estimating Yaw in magnetically disturbed environments
The gyroscope bias is continuously estimated. For the rate of turn around the x-axis and the y-axis (roll and pitch axes), the gyroscope bias is estimated using gravity (accelerometers). In a homogenous magnetic field and with filter profiles using the magnetometer, the gyroscope bias around the z-axis will successfully be estimated.
In some situations, the heading cannot be referenced to the (magnetic) north. This is the case when the magnetic field is not used (for example for VRU/AHT devices) or when the magnetic field is not homogenous. There are several ways to mitigate the drift in heading (rotation around the z-axis):
The XKF3i algorithm not only computes orientation, but also keeps track of variables such as sensor biases or properties of the local magnetic field (magnetic field: AHRS only). For this reason, the orientation output may need some time to stabilize once the MTi is put into measurement mode. Time to obtain optimal stable output depends on a number of factors. An important factor determining stabilizing time is determined by the time to correct for small errors in the bias of the rate gyroscopes. The bias of the rate gyroscope may slowly change due to different effects such as temperature change or exposure to impact.
As described above, XKF3i uses assumptions about the acceleration and the magnetic field to obtain orientation. Because the characteristics of the acceleration or magnetic field differ for different applications, XKF3i makes use of filter profiles to be able to use the correct assumptions given the application. This way, XKF3i can be optimized for different types of movement. For optimal performance in a given application, the correct filter profile must be set by the user. For information on how to specify a filter profile in XKF3i, please refer to the MT Manager User manual or the MT low-level communication protocol documentation.
Filter profiles for the MTi-200/MTi-300
|
Number |
Name |
IMU |
Magnetometer |
Product |
|
39 |
General |
• |
• |
30/300-AHRS |
|
40 |
High_mag_dep |
• |
• |
30/300-AHRS |
|
41 |
Dynamic |
• |
• |
30/300-AHRS |
|
42 |
Low_mag_dep |
• |
• |
30/300-AHRS |
|
43 |
VRU_general |
• |
|
30/300-AHRS; 20/200-VRU/AHT |
The general filter profile is the default setting. It assumes moderate dynamics and a homogenous magnetic field. External magnetic distortions are considered relatively short (up to ~20 seconds). Typical applications include camera tracking (e.g. TV camera’s), remotely operated robotic arms on ROV’s etc
The high_mag_dep filter profile assumes a homogenous magnetic field and an excellent Magnetic Field Mapping. This filter profile heavily relies on the magnetometer for heading. Dynamics are relatively slow. Typical applications are navigation of ROV’s or the control of small unmanned helicopters.
The dynamic filter profile assumes jerky motions. However, the assumption is also made that there is no GNSS available and/or that the velocity is not very high. In these conditions, a 100-series MTi may be a better choice. The dynamic filter profile uses the magnetometer for stabilization of the heading, and assumes very short magnetic distortions. Typical applications are where the MTi is mounted on a person or hand-held (e.g. HMD, sports attributes etc.).
The low mag_dep filter profile assumes that the dynamics is relatively low and that there are long-lasting external magnetic distortions. Also, use this filter profile when it is difficult to carry out a very good Magnetic Field Mapping (MFM). The use of the low_mag_dep filter profile can be useful to limit drift in heading whilst not being in a homogenous magnetic field. Typical applications are large vessels and unmanned ground vehicles in buildings.
The VRU_general filter profile assumes moderate dynamics in a field where the magnetic field cannot be trusted at all and benefits from the Active Heading Stabilization feature. It is also possible to use this filter profile in situations where an alternative source of yaw is available. Yaw from the VRU/AHT is unreferenced; note however, that because of the working principle of the VRU/AHT, the drift in yaw will be much lower than when gyroscope signals are integrated only. Typical applications are stabilized antenna platforms mounted on cars of ships and pipeline inspection tools. This filter profile is the only one available for the MTi-20/200 VRU/AHT.
Every application is different and although example applications are listed above, results may vary from setup to setup. It is recommended to reprocess recorded data with different filter profiles in MT Manager to determine the best results in your specific application.
The Xsens sensor fusion algorithm in the MTi-G-710 has several advanced features. It can handle a multitude of data channels, to incorporate GNSS and barometer data as well.
The MTi-G-710 algorithm adds robustness to the orientation and position estimates by combining measurements and estimates from the inertial sensors and GNSS receiver in order to compensate for transient accelerations. This results in improved estimates of roll, pitch, yaw and position.
Next to the solutions described on BASE to mitigate effects of magnetic disturbances, the MTi-G-710 sensor fusion algorithm makes use of data coming from the GNSS receiver. This means that the MTi-G-710 has an increased resistance to magnetic disturbances. It is, for example, possible to estimate the heading based on comparison between accelerometer data and the GNSS acceleration.
When the MTi-G-710 has limited/mediocre GNSS reception or even no GNSS reception at all, the MTi-G-710 sensor fusion algorithm seamlessly adjusts the filter settings in such a way that the highest possible accuracy is maintained. The sensor will continue to output position, velocity and orientation estimates, although the accuracy is likely to degrade over time as the filters will have to rely on dead-reckoning. The GNSS status will be monitored continuously so that the filter can take GNSS data into account when available and sufficiently trustworthy. In case the loss of GNSS lasts longer than a period of 45 seconds, the MTi-G-710 will go into a state where it stops producing position and velocity estimates, and no longer uses velocity estimates in its sensor fusion algorithms until GNSS reception is re-established. An exception to this is the HighPerformanceEDR filter profile, which performs extended dead-reckoning up to 600 seconds.
The filter profiles for MTi-G-710 are described below. Please note the specific cautions with each of these filter profiles.
Filter profiles for the MTi-G-710 GNSS/INS
|
Nr |
Name |
IMU |
Mag field |
Static pressure |
GNSS |
Holonomic constraints |
Product |
|
1 |
General |
• |
|
• |
• |
|
710-GNSS/INS |
|
2 |
GeneralNoBaro |
• |
|
|
• |
|
710-GNSS/INS |
|
3 |
GeneralMag |
• |
• |
• |
• |
|
710-GNSS/INS |
|
4 |
Automotive |
• |
|
• |
• |
• |
710-GNSS/INS |
|
9 |
HighPerformanceEDR |
• |
|
• |
• |
|
710-GNSS/INS |
The General filter profile is the default setting. It makes few assumptions about movements. Yaw is referenced by comparing GNSS acceleration with the on-board accelerometers, so more movement (when GNSS is available) will result in a better yaw. Altitude (height) is determined by combining static pressure, GNSS altitude and accelerometers. The barometric baseline is referenced by GNSS, so during GNSS outages, accurate height measurements are maintained because this barometric baseline is monitored.
The GeneralNoBaro filter profile is very similar to the general filter profile. However, it does not use the barometer for height estimation (it thus uses GNSS and accelerometers only). Since airflow near the venting holes in the MTi-G will lower the barometric pressure (and thus make height estimations inaccurate), you can use this filter profile when the MTi-G is mounted in such airflow.
The GeneralMag filter profile bases its yaw mainly on magnetic heading, together with comparison of GNSS acceleration and the accelerometers. Although this combination makes the yaw more robust than the magnetic field alone, a homogenous or calibrated-for magnetic field is essential for good-performance yaw. Other parameters are tuned the same as in the General filter profile.
The Automotive filter profile assumes that the yaw of the MTi-G-710 is also the GNSS course over ground (holonomic constraints). This assumption holds for most automotive/ground vehicles, except for those who experience side slip, such as racing cars, tracked vehicles, some articulated vehicles (depending on where the MTi-G-710 is mounted) and vehicles driving on rough terrain. The Automotive filter profile thus uses GNSS to determine the yaw. Note that it is essential to mount the MTi-G-710 exactly in the direction of movement in order to prevent an offset. When GNSS is lost, yaw will be determined by the velocity estimation algorithm for 45 seconds, before yaw is determined by gyroscope integration only. Should GNSS outages recur regularly or if you have bad GNSS-availability (e.g. in urban canyons), consider using HighPerformanceEDR.
The HighPerformanceEDR filter profile replaces the previously available AutomotiveUrbanCanyon filter profile. This filter profile is specially designed for ground-based navigation applications where deteriorated GNSS conditions and GNSS outages (0-600s) are a regular feature. Note that the accuracy of position, velocity and orientation estimates may still deteriorate during GNSS outages. This filter profile does not use the holonomic constraints and thereby removes the need for mounting considerations. Target applications: slow moving ground vehicles and locomotive navigation. The filter profile HighPerformanceEDR automatically estimates the gyro bias when the MTi is not moving. The sensor fusion algorithm detects when the MTi is motionless. Vibrations and very slow movements may influence the accuracy of the gyro bias estimation.
Every application is different and although example applications are listed above, results may vary from setup to setup. It is recommended to reprocess recorded data with different filter profiles in MT Manager to determine the best results in your specific application.
u-blox receivers support different dynamic platform models in order to adjust the navigation engine to the expected application environment. The MTi-G-710 can be configured to use a desired platform model upon start-up. This enables the user to adjust the u-blox receiver platform to match the dynamics of an application. The setting influences the estimates of Position and Velocity and therefore, it affects the behaviour of the Xsens filter output.
The platform model can be configured using MT Manager or low-level communication. Currently, MT Manager only allows you to select the Portable (default) or Airborne (<4g) platform, but more options are available using low-level communication by providing the GNSS Platform ID. For more details on GNSS platform settings, refer to the u-blox Receiver Description Manual.
The Orientation Smoother is a software component within the GNSS/INS sensor fusion engine. This feature aims to reduce any sudden jumps in the Orientation outputs that may arise when fusing low-rate GNSS receiver messages with high-rate inertial sensor data.
The Orientation Smoother can be enabled from the Device Settings window in MT Manager, or by using the setOptionFlags low-level command (see MT Low Level Communication Protocol Documentation).
Out of the filter profiles available for the MTi-G-710, only the GeneralMag filter profile will provide the user with a North-referenced heading when the MTi is powered up. For all other filter profiles, the Yaw will initialize at 0 degrees. As soon as a GNSS fix is available and the MTi starts moving at a sufficient velocity, the Yaw will converge to a North-referenced heading.
If the initial heading of the MTi is approximately known at initialization time, for instance, based on an external sensor or a known starting orientation, then it is possible to manually set the initial heading value. This can be done using the SetInitialHeading low-level communication command. Refer to the MT Low Level Communication Protocol Documentation for more information.
One powerful feature when it comes to heading estimation is Active Heading stabilization (AHS). The AHS is a software component within the sensor fusion engine designed to give a low-drift unreferenced Yaw solution. It uses the magnetic field to stabilize the Yaw output. This way, drift in Yaw can be as low as 1° after 60 minutes for the MTi 100-series and 3° after 60 minutes for the MTi 10-series. Even if the magnetic field is disturbed, AHS will still function.
AHS is not tuned for nor intended to be used with GNSS/INS devices.. Therefore, Xsens discourages the use of this feature for GNSS/INS devices, such as the MTi-7 and MTi-G-710.
For more information on the activation and use of AHS, refer to the BASE: Active Heading Stabilization
When it is not possible to carry out a Magnetic Field Mapping and when there are hard- and soft-iron effects that are moving with the MTi, it is possible to perform an In-run Compass Calibration (ICC).
In-run Compass Calibration is a way to calibrate for magnetic distortions that move with the sensor using an onboard algorithm, leaving out the need for a host processor like a PC. However, if possible, using the Magnetic Field Mapper tool is preferred. It estimates the hard- and soft-iron effects and provides new magnetometer calibration parameters. To expedite the estimation of magnetometer calibration parameters, there is a Representative Motion feature. Representative Motion is available in MT Manager, XDA and Low-Level Communication Protocol (Xbus protocol).
ICC is currently a feature in beta. For more information, refer to BASE: In-Run Compass Calibration
High quality sensor input is key for good orientation and position performance in an application. Therefore, it is important to take care when installing the sensor (see also following chapters) as well as allow the internal algorithms to have a good estimation of sensor parameters.
Even though during Xsens factory calibration all characteristics of each sensing element are written in the memory of the MTi, there are some characteristics, for example, gyroscope bias, which can change over the lifetime or per power cycle of the MTi device. A more in-depth discussion on sensor bias can be found on BASE: Understanding Sensor Bias
During the initialization phase, the performance of the MTi may be sub-optimal, because the gyroscope biases are not yet estimated correctly. This chapter is to provide two best practices to allow the algorithm to reach optimal performance as quickly as possible by allowing it to estimate the gyroscope biases.
The internal algorithms of the MTi can estimate the biases of the gyroscopes while it is in use. However, upon start-up, there has not yet been a possibility to estimate the biases properly. The Manual Gyro Bias Estimation functionality is designed to allow the MTi to estimate the gyro biases quickly (<10s) and accurately.
A typical application start-up procedure could be as follows:
More information on the use of Manual Gyro Bias Estimation can be found on BASE: Manual Gyro Bias Estimation
The MTi can use the direction of gravity to determine its gyroscope biases. The MTi uses three gyroscopes in perpendicular axes. This means that two gyroscope axes are perpendicular to gravity and one gyroscope axis is aligned with gravity when the MTi is placed on a horizontal surface.
In this situation, the bias of the gyroscope that is aligned with gravity cannot be estimated. Therefore, if 3D motion is applied to the sensor, e.g. the vertical axis is rotated sufficiently away from the vertical, the last sensor bias becomes observable to the filter of the MTi.