Magnetic Sensor Compensation Utilizing Factor Graph Estimation
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Electrical Engineering. The day in electrical engineering: signals, systems and the hardware that carries them.
Rosa: I'm Rosa.
Dev: And I'm Dev, here with Taro.
Rosa: We're looking at the paper: "Magnetic Sensor Compensation Utilizing Factor Graph Estimation".
Dev: Here's the gist of the paper. Magnetic calibration algorithms based on factor graph optimization incorporating inertial measurements are presented to accurately calibrate magnetic sensors in flight, addressing challenges related to large permanent moments and time-varying external magnetic fields <ref:2312.14343#pg9>.
Rosa: This method offers greater flexibility for sensor mounting locations in magnetic anomaly navigation <ref:2312.14343#pg3>.
Dev: Attempting to utilize magnetic anomaly signals for navigation has been researched for several decades <ref:2312.14343#pg2>.
Rosa: A key technical bottleneck to achieving magnav is obtaining an accurate magnetic sensor calibration, taking into account own-ship and sensor effects <ref:2312.14343#pg3>.
Paper summary: Rosa: They're moving away from just using models for permanent moments and eddy currents, which is what Tolles-Lawson does. Instead, they use a factor graph that can handle vector and scalar sensors at the same time >
Taro: Can you tell us more about how this factor graph handles those different sensor types? I mean, the paper mentions it's well suited for combining multiple sensor inputs >
Dev: It seems like the state of their system includes things like hard iron, vector bias, and scale factor errors along with orientation at each time step >
Rosa: Exactly. They model the magnetic measurement in a few ways: one for vector measurements where they have an equation involving T, scale factor, ortho error, nonorthogonality error sigma and bias error plus hard iron >
Taro: And for scalar sensors, the model is simpler because it doesn't have that vector component >
Dev: The paper shows that this factor graph calibration method still performs better than methods like TWOSTEP and Tolles-Lawson when there are hard iron contributions in the simulations >
Conclusion: Rosa: So, to wrap up this paper by Lathrop, Taylor, and Nielsen, they developed this magnetic calibration algorithm using factor graph optimization combined with inertial measurements >
Dev: The real implication here is that this approach is much more robust against big hard iron effects than the prior methods they compared it to >
Taro: It also allows for calibrations in environments where you can't control everything perfectly, like a real flight environment with a time-varying external magnetic field >
Rosa: That flexibility regarding sensor placement and handling those varying fields is what makes this method potentially useful for practical applications in magnetic anomaly navigation >
Air Force Institute of Technology
eess.SP
Submitted: 2023-12-22
Updated: 2023-12-22
Journal ref: IEEE Sensors Journal, 24(15), 2024
DOI: 10.1109/JSEN.2024.3416618
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 72/100
The gist: Magnetic calibration algorithms based on factor graph optimization incorporating inertial measurements are presented to accurately calibrate magnetic sensors in flight…
Key concepts
- Factor Graph Framework
- A flexible optimization technique used for complex, non-linear problems like magnetic calibration. It allows the system to model and solve relationships between different measurements simultaneously. This framework enables the calibration of both vector and scalar sensors in a unified manner.
- Magnetic Measurement Model
- This mathematical model describes how magnetic sensor readings are affected by various errors, including sensor bias, hard iron distortion, soft iron distortion, non-orthogonality, scale factor errors, and noise. It is used to accurately relate the actual magnetic field to the measured data.
- Hard Iron Effects
- These refer to the magnetic distortions caused by the physical structure of the aircraft or platform itself. The paper focuses on improving calibration methods that account for these large permanent moments, showing its robustness in handling significant hard iron contributions compared to older techniques.
Terminology
Summary
Magnetic calibration algorithms based on factor graph optimization incorporating inertial measurements are presented to accurately calibrate magnetic sensors in flight, addressing challenges related to large permanent moments and time-varying external magnetic fields <ref:2312. p. 114343#pg9>. This method offers greater flexibility for sensor mounting locations in magnetic anomaly navigation <ref:2312. p. 114343#pg3>.
Introduction to the Problem
Attempting to utilize magnetic anomaly signals for navigation has been researched for several decades <ref:2312.14343#pg2>. A key technical bottleneck to achieving magnav is obtaining an accurate magnetic sensor calibration, taking into account own-ship and sensor effects <ref:2312. p. 114343#pg3>. The Tolles-Lawson magnetic calibration method is the foundation still used today for platform compensation for most aerial magnetic surveys <ref:2312. p. 214343#pg7>. This method uses an aircraft magnetic disturbance model that identifies the three primary types of disturbances: permanent moments, induced moments, and eddy currents <ref:2312. p. 214343#pg8>.
The Factor Graph Framework
Factor graphs present a flexible optimization method suitable for non-linear problems such as magnetic calibrations <ref:2312.14343#pg9>. The flexibility also allows the factor graph to be constructed such that it calibrates both vector and scalar sensors, unlike previously described methods that calibrate either a scalar or vector sensor, but not both <ref:2312. p. 214343#pg9>. The state of the system is represented by a vector including hard iron (hhi), vector bias (hvec), scale factor and nonorthogonality errors (T), and the orientation at each time step <ref:2312. p. 614343#pg10>.
Calibration Model Details
The magnetic measurement model combines sensor errors such as sensor bias, hard iron distortion, soft iron distortion, non-orthogonality, scale factor, and noise <ref:2312.14343#pg6>. The vector magnetic measurement is modeled by the equation mvec = T scale T ortho T si C b ne n + hhi + hvec + ϵvec <ref:2312. p. 414343#pg7>. The scalar sensor model is simplified to mscalar = T si C b ne n + hhi + ϵscalar <ref:2312. p. 614343#pg9>.
Performance and Results
The factor graph calibration method was tested with both simulated and real data sets <ref:2312.14343#pg10>. The results show that the Factor Graph calibration still significantly outperformed TWOSTEP and Tolles-Lawson in simulations with hard iron contributions <ref:2312. p. 314343#pg15>. In experimental evaluation, the Factor Graph method had the lowest error when using the same measurements to estimate the hard iron characteristics of the platform <ref:2312. p. 314343#pg20>. The paper concludes that this algorithm is significantly more robust to large hard iron effects than prior approaches <ref:2312. p. 114343#pg20>.
Limitations and Future Work
The primary shortcoming to address is the current algorithm was limited to calibrating hard iron effects, whereas other contemporary routines solve for both hard iron and soft iron, and in the case of Tolles-Lawson, eddy currents <ref:2312.14343#pg20>. Additionally, it is assumed that the hard iron effects are the same for both the scalar and vector sensors <ref:2312. p. 514343#pg20>. The ability to allow a time-variant external field allows calibrations in environments that cannot be as tightly controlled as a lab experiment <ref:2312. p. 114343#pg20>.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improve magnetic sensor calibration by utilizing a factor graph optimization that incorporates both inertial measurements and inputs from both a vector and scalar magnetometer to achieve accurate calibrations in flight. This allows for
greater flexibility in sensor mounting locations for magnetic anomaly navigation.
-
Enable magnetic anomaly navigation systems to accurately calibrate the platform even in the presence of large permanent moments, which is difficult with previous approaches. The method can
accurately calibrate a magnetic sensor in flight (i.e., with the presence of a large platform field and a varying Earth field due to movement).
-
Develop an AI system capable of estimating a time-varying external magnetic field during calibration while simultaneously calibrating the sensor, moving beyond methods that assume a fixed external field. This capability is demonstrated by the Factor Graph method's ability to handle
a time varying external magnetic field during the calibration.
-
Enhance navigation systems to improve performance when hard iron effects are large, as shown by the paper's results demonstrating that
the Factor Graph calibration still significantly outperformed TWOSTEP and Tolles-Lawson in simulations with hard iron contributions.
Sources
- Derivation and Extensions of the Tolles-Lawson Model for Aeromagnetic Compensation
- Online 3-Axis Magnetometer Hard-Iron and Soft-Iron Bias and Angular Velocity Sensor Bias Estimation Using Angular Velocity Sensors for Improved Dynamic Heading Accuracy
Related papers
- Open-H-RF: A Large-Scale Ultrasound Raw Channel Dataset Initiative
- Scalable Attention for 5G NR Channel Estimation
- Low-Altitude Wireless Networks: A Comprehensive Survey
- Sensing-Assisted OFDM with Index Modulation: A Waveform for Optical Wireless Integrated Communication and Localisation
- Iteration-Free Candidate Pruning and an Index-Domain VLSI Architecture for Sort-Free MIMO Box Decoding up to 1024-QAM
- A Foundation Model for Large-Scale Wireless Network Planning, Operation and Optimization