Tightly Coupled Navigation System of a Differential Magnetometer System with a MEMS-IMU for Enceladus
Sabine Macht, Martin Escher, Markus Bobbe, Barbara Kohn, Ulf Bestmann, Institue of Flight Guidance, Germany
Location: Big Sur
Date/Time: Thursday, Apr. 26, 8:35 a.m.
Navigation under ice is a special challenge since most optical and radio frequency based systems are not able to penetrate the ice with reasonable range. To get further information of the genesis of life in our galaxy, exploration of the surface and shell of various iced planets or their moons is of interest. One is Saturn’s moon Enceladus, where some of life constitution conditions are met.
In the „EnEx Enceladus Explorer“-project funded by „DLR Raumfahrtmanagement“ an automatic navigating in-ice probe was developed to get samples of long isolated water containing different ancient bacteria. It was equipped with different navigation technologies, e.g. inertial sensors, acoustic positioning, reconnaissance systems and especially a differential magnetometer system. For navigation a combination of dead reckoning and acoustic ranging has been developed to combine the benefits of both systems. Using this system successful taking samples in glaciers is demonstrated at the South Pole. This project ended in 2015. Meanwhile, the probe and its navigation technologies were developed further.
In order to get a better environment reconnaissance and advance navigation quality a follow-up project to the first EnEx project was established. Its main goal is to identify various obstacles by differential magnetic measurements and to enhance navigation quality as well. In addition, a miniaturized MEMS-InertialMeasurement Unit was installed to minimize space and power consumption. This exchange of measurement equipment is followed by some changes in navigation and reconnaissance algorithms.
This paper focuses on the development and test of the advanced magnetic navigation and reconnaissance systems inside the IceMole navigation system, which enables precise probe navigation in the ice. Navigation under ice is a special application of borehole navigation, where measuring the drilled depth and driven distance is difficult. However, results of magnetic support are widely influenced by the quality of magnetic calibration.
Due to slow forward motion and low rates common strapdown systems would not work well inside the ice probe. Especially a MEMS-IMU’s gyros show a rising drift and a small signal-to-noise ratio caused by the missing external stimulation. This is compensated by an acceleration-only mode for the IMU measurements in the filter system. Further, the motion behavior of the IceMole is constrained inside the melting path. Rotation of the probe is given by re-actio forces induced by the ice-screw. These forces depend on heating power and ice-screw turn rate as well as glacier temperature and density. The actual angle of roll and its change are a-priori unknown. Furthermore a sudden pitching of the probe is not possible as long as the probe stays outside crevasses.
These constrains yield to an acceleration-only, tightly-coupled navigation filter, where a 2-dimensional attitude is given by the measured and calculated gravity vector. Arising singularities caused by the possible 360 degree roll of the IceMole are handled by using quaternions in navigation calculation. To get a full navigation solution and to compensate missing rotation information a differential magnetometer system is investigated. Especially heading and bank support are needed as aiding input. The reviewed differential magnetometers are the same sensor type as the magnetometers used in the former EnEx-Project and real space missions like ROSETTA.
Using magnetometers means having a magnetic influence by surrounding technical equipment like computers, converters, heaters, power train and sonic reconnaissance systems in measurements. Even the magnetic sensor oneself have an electro-magnetic interaction when mounted close-by. These disturbers influence the detection of motion. In order to get a reliable heading solution and to differentiate (magnetic) obstacles from the surrounding, magnetic calibrations have to be executed. Calibrations of magnetic sensors include different tests and an approximation of the measured value afterwards to gain a full rotation matrix for later online corrections. The measured values are best-fit to a sphere whose magnitude is given by local magnetic flux. This best-fit correction is calculated by the Rodrigues-formula, whose main goal is a linear transformation by axis and angle. However, a sensor installed near heaters or drive is driven into saturation which can only be detected but not compensated. Measuring values in sensor-saturation should be rejected by the coupling filter. Heading calculation has to include magnetic corrections as well as filtering for noise reduction. All the same, real time conditions will still be met due to the slow movement of the melting probe.
A challenge will be the transfer and adaption of the navigation system, whose development take place in earth‘s conditions, to the variable Saturn’s magnetic field at Enceladus. Hence, differential magnetic algorithms for navigation on Saturn’s moon Enceladus were a subject of investigation as well. Saturn’s magnetic field measured by Cassini provides a basis for a detailed magnetic-field simulation. While having no magnetic field of its own, Enceladus is moving in Saturn’s magnetosphere. To get a usable magnetic source for navigation, a-priori knowledge of orientation and strength of the expectable magnetic field at the target area should be gained: knowing strength and orientation as well as variation of the magnetic field Enceladus is moving in, an estimation of the navigation accuracy and precision of an autonomous navigation probe will be easier. During the EnEx-MIE project a simulation was build that calculates Enceladus’ orbit around Saturn and moves the position of measurement to the moon’s south pole, where the so-called “Tiger Stripes” were seen by Cassini. These stripes are the target area for the proposed mission since they are supposed to be connected to the subglacial ocean at Enceladus and are of interest to astrobiologists. In conclusion the earth-based magnetic navigation solution will be transferred to space requirements for a later Enceladus Mission which may take place in some years’ time.
The paper further shows the performance of the developed algorithms using data of Field tests from the Langenferner Gletscher at the Ortles Alps in Italy and simulation results for the moon Enceladus.