A Heptafurcated Optical-Fiber Bundle for Simultaneous Distance, Tilt Magnitude and Tilt-Axis Sensing

Technical Summary: A Heptafurcated Optical-Fiber Bundle for Simultaneous Distance, Tilt Magnitude and Tilt-Axis Sensing

Problem Statement
Reflective intensity-modulated optical-fiber sensors are widely used for non-contact metrology due to their compactness, passivity, and robustness. However, conventional fiber-bundle architectures typically measure either axial distance or angular displacement in isolation. They generally fail to simultaneously recover the three geometric parameters required to fully characterize a target's position and orientation: axial distance ($d$), tilt magnitude ($\alpha$), and tilt-axis azimuth ($\Phi$). While various multi-axis sensors exist (e.g., interferometric, FBG-based, or grating-reflector systems), they often rely on complex interrogation structures or do not utilize a single passive intensity-based bundle to solve for all three parameters simultaneously.

Methodology
The authors propose a novel Heptafurcated Optical-Fiber Linear and Angular Displacement Sensor (OFLADS). The sensing head consists of a single passive bundle containing:

1. One central transmitting fiber (single-mode).
2. Four directional receiving arms arranged along the $+x, -x, +y,$ and $-y$ axes.
3. Two concentric receiving rings (inner and outer).

This configuration yields six distinct power measurements ($P_{+x}, P_{-x}, P_{+y}, P_{-y}, P_{in}, P_{out}$).

Mathematical Modeling and Reconstruction Strategy
A complete forward model was developed to describe the optical coupling, accounting for:

• Gaussian beam emission and specular reflection on a tilted mirror.
• Finite-core coupling effects.
• Numerical-aperture (NA) attenuation.
• Mechanical admissibility (preventing bundle-tip contact with the target).
• An unknown global gain factor ($G$) representing source power, reflectivity, and electronic gain.

The reconstruction procedure follows a three-stage physically structured approach to solve the inverse problem:

1. Azimuth Estimation: The tilt-axis azimuth ($\Phi$) is estimated directly from the logarithmic ratios of opposite arms ($L_x = \ln(P_{+x}/P_{-x})$ and $L_y = \ln(P_{+y}/P_{-y})$). The azimuth is calculated as $\hat{\Phi}_0 = \text{atan2}(L_x, -L_y)$.
2. Initial Distance and Tilt Estimation: Once $\Phi$ is known, the arm responses are projected along the tilt direction ($L_\parallel$) and combined with the ring log-ratio ($Q = \ln(P_{out}/P_{in})$). These observables are mapped against a pre-calibrated lookup table to obtain initial estimates for distance ($d_0$) and tilt magnitude ($\alpha_0$).
3. Joint Refinement: A nonlinear inversion is performed using the full six-power vector. The unknown common gain is eliminated analytically, and the parameters $(d, \alpha, \Phi)$ are jointly refined to minimize the weighted logarithmic residual between measured and modeled powers.

Experimental Validation
The sensor was validated using a setup comprising a 660 nm laser, a flat mirror target, and precision linear/angular stages. Measurements were taken at five representative operating conditions spanning distances of 2–9 mm and tilt magnitudes of 3–8 degrees. The system utilized gain-equalization calibration to mitigate channel-to-channel variations.

Key Results
The experimental validation demonstrated the sensor's ability to simultaneously reconstruct all three parameters with the following performance metrics:

• Distance ($d$): Mean absolute error of 1.63%; maximum absolute error of 5.44%.
• Tilt Magnitude ($\alpha$): Mean absolute error of 2.06%; maximum absolute error of 4.65%.
• Tilt-Axis Azimuth ($\Phi$): Mean absolute error of 0.87°; maximum absolute error of 1.74°.

The analysis revealed that the cross-arms provide directional information (sensitive to $\Phi$), while the concentric rings provide radial information (sensitive to $d$ and $|\alpha|$ but largely independent of $\Phi$). This separation of roles is critical for resolving the coupling between distance and tilt. The authors noted that performance degrades slightly in regions of large distance and high tilt due to weaker conditioning of the inverse problem.

Significance and Claims
The paper claims that the OFLADS offers a compact, passive, and potentially low-cost solution for simultaneous three-parameter non-contact monitoring. Its primary significance lies in:

1. Simultaneity: It recovers distance, tilt magnitude, and tilt-axis orientation from a single bundle using only scalar intensity measurements, a capability not found in conventional bundle OFDSs.
2. Architecture: The specific "heptafurcated" geometry (four arms + two rings) successfully encodes complementary directional and radial information, enabling the separation of the tilt-axis azimuth from the coupled distance-tilt problem.
3. Applicability: The authors suggest the sensor is suitable for industrial applications such as positioning, alignment, and robotic metrology where the target cannot be instrumented directly and where space or environmental constraints (vibration, EMI) favor passive optical solutions.

The work does not claim to outperform interferometric or spectral methods in absolute precision but emphasizes the unique ability to achieve multi-parameter sensing with a simple, robust, and single-bundle intensity-modulated architecture.