Grating Wheel Assembly (GWA)

• Grating Wheel Assembly (GWA) is a precision electromechanical mechanism that rotates and positions multiple optical elements for spectroscopy and adaptive terrain applications.
• It integrates high-resolution sensors such as dual magneto-resistive units and encoder systems to achieve sub-microradian angular precision through real-time feedback.
• Its design employs robust calibration models and adaptive control to mitigate mechanical errors and environmental effects, ensuring optimal performance in both space-based and terrestrial platforms.

A Grating Wheel Assembly (GWA) is a precision electromechanical mechanism designed to select among multiple optical elements—most typically diffraction gratings, prisms, and mirrors—by rotating them into the optical path of a spectrometer system. GWAs have been implemented both in space-based instrumentation, most prominently in the Near-Infrared Spectrograph (NIRSpec) on the James Webb Space Telescope (JWST), and in terrestrial or planetary mobility platforms requiring high-fidelity control of traction via wheel geometry. GWAs operate under stringent requirements for positional repeatability, thermal stability, and reliable real-time sensing or feedback, as their functional accuracy directly determines system-level performance in spectral calibration, target acquisition, or terrain adaptation (Oliveira et al., 2022, Griffo et al., 23 May 2026).

1. Mechanical and Optical Architecture

In the context of space instrumentation, the GWA of JWST/NIRSpec consists of a cryogenic rotating wheel mechanism housing eight optical elements: six diffraction gratings (split into three medium-resolution, $R \approx 1000$, and three high-resolution, $R \approx 2700$), a R~100 double-pass prism, and a flat mirror for imaging and target acquisition (Oliveira et al., 2022). The wheel is mounted on a stainless-steel ball-bearing mechanism operating at $\sim$37 K. Actuation is provided by a cryo-rated brushless torque motor, with a spring-loaded ratchet ("click wheel") for kinematic seating in eight detents, and isostatic mounts minimizing deformation and stresses during thermal cycling.

For adaptive planetary mobility, the GWA refers to a drive wheel with actively variable grouser height, actuated by a high-torque servo motor through a spiral-cam mechanism. The mechanical subsystem includes a central servo driving an internal planetary gearbox with a $-7.5{:}1$ reduction, and a separate gearmotor for wheel propulsion. The spiral-cam (stroke $0$–$17.5$ mm) is implemented in 6061-T6 aluminum, its profile optimized for minimal pressure angle (maximum $25^\circ$) to reduce torque. Sixteen microbearing-guided followers ensure uniform vertical displacement of all grousers. Structural dimensions, such as wheel radius ($r=62.5$ mm), width (42 mm), and grouser width (25 mm, chamfer $45^\circ$), are fixed (Griffo et al., 23 May 2026).

2. Sensor Systems and Feedback

High-precision angular positioning in spectrograph GWAs is achieved via dual magneto-resistive (MR) sensors. Each MR sensor is sensitive to the fringe field from two permanent magnets fixed to the wheel, measuring tip/tilt axes: one along the dispersion ("spatial") axis, the other along the cross-dispersion ("spectral") axis. The output voltages $(V_1, V_2)$ form a unique signature for the angular displacement of the selected optic. Mechanical repeatability of the uncorrected wheel is limited to $R \approx 2700$0–$R \approx 2700$1rad ($R \approx 2700$2–$R \approx 2700$3 detector pixels), but sensor resolution extends to $R \approx 2700$4rad, with in-situ calibration reducing the effective wheel-induced image motion to $R \approx 2700$5rad (well below the science-level threshold) (Oliveira et al., 2022).

For adaptive robotic wheels, feedback is provided via:

• a DC motor encoder (wheel position/velocity),
• an AS5600 magnetic encoder (cam angle),
• a 5 $R \approx 2700$6m linear encoder (for actual platform displacement velocity), and
• an INA169 current sensor (motor current/power) (Griffo et al., 23 May 2026).

Closed-loop height regulation is implemented by mapping cam angle $R \approx 2700$7 to grouser height $R \approx 2700$8 using a piecewise cubic spline, enabling real-time PID control to maintain $R \approx 2700$9 with $\sim$0 mm precision under load.

3. Calibration and Mathematical Modeling

NIRSpec GWA sensors require rigorous calibration of their voltage-to-angle mapping. The two-axis model is:

$\sim$1

This affine transformation incorporates systematic offsets and cross-coupling terms (non-orthogonality) for each optical element; individual calibrations are performed both in cryo-vacuum and in-flight conditions. The mapping shows $\sim$2 linearity for all elements, with no meaningful coefficient drift between ground and in-flight calibration at the $\sim$3 level (Oliveira et al., 2022).

For terrain-adaptive GWAs, the cam angle–height function is derived in CAD and realized via spline interpolation. Closed-loop height error is computed as $\sim$4, with continuous PID feedback as:

$\sim$5

Metrics critical for mobility (slip ratio, energy, and traction force) are derived via real-time sensing:

• Slip ratio:

$\sim$6

• Energy:

$\sim$7

• Force (first approximation):

$\sim$8

Grouser geometry–terrain compatibility is assessed with the Inotsume et al. criterion for non-interfering grouser spacing.

4. Operational Performance and Integration

For NIRSpec, each time a new GWA element is selected, the two MR sensors are read out and the angular displacements are inserted into the NIRSpec instrument model (Lutzgendorf et al. 2022). This corrects spatial and spectral distortions in real-time for both on-board target-acquisition and ground data reduction. Residual uncertainty after calibration ($\sim$9rad) corresponds to $-7.5{:}1$0 pixel on the detector—satisfying all operational requirements for multi-object spectroscopy and precise target placement (Oliveira et al., 2022).

In terrain-adaptive robotic applications, on-the-fly grouser actuation permits real-time adjustment to variable terrain surface properties. In 750 experimental trials across four terrains (vinyl sheet, coarse rock, pea gravel, fine sand in two states), adaptive deployment of the grouser array reduced slip by $-7.5{:}1$1–$-7.5{:}1$2 and improved travel time and energy usage by up to $-7.5{:}1$3 over fixed configurations. No single fixed grouser height minimized slip across all terrains, reinforcing the operational advantage of continuous adaptation (Griffo et al., 23 May 2026).

5. Scaling Laws and Design Guidelines

An empirically derived relationship connects terrain median particle diameter $-7.5{:}1$4 (in mm) to optimal grouser height $-7.5{:}1$5:

$-7.5{:}1$6

Smaller $-7.5{:}1$7 (finer grains) require larger $-7.5{:}1$8. For a $-7.5{:}1$9 mm-diameter wheel at $0$0 kg load under Earth's gravity, recommended optima are $0$1 mm (hard/rigid), $0$2 mm (mid-coarse), and $0$3 mm (fine sand). However, measured field conditions—specifically particle angularity, packing state, normal load, and gravity—can shift these optima, necessitating in situ validation (Griffo et al., 23 May 2026). Closed-loop slip feedback is essential for robust execution under variable conditions.

6. Performance Metrics, Residual Errors, and Limitations

The NIRSpec GWA sensor system achieves post-calibration residuals of $0$4–$0$5rad (equivalent to $0$6), with negligible drift between ground and space operation. The system's error budget is dominated pre-calibration by wheel bearing repeatability ($0$7–$0$8rad) but post-calibration by electronic noise ($0$9rad) and residual model fitting error ($17.5$0rad). This level of control ensures the GWA is not a limiting factor for either wavelength or spatial addressing in science operations (Oliveira et al., 2022).

For adaptive terrain wheels, slip, travel time, and energy consumption are all metrics directly measured and minimized. Predicted optimal grouser heights yielded incremental slip reductions of $17.5$1–$17.5$2 over fixed best cases for coarse terrains; in sand, differences were $17.5$3. These results validate the core utility of real-time morphing geometry for planetary or terrestrial mobility (Griffo et al., 23 May 2026).

7. Broader Implications and Engineering Lessons

The GWA demonstrates the utility of integrating high-resolution, robust sensor feedback with a simple, physically-motivated calibration model. In the spectroscopy domain, a two-axis linear model suffices to decorrelate wheel-induced distortions from science data, obviating the need for more complex or adaptive inference. In mobility systems, the GWA paradigm illustrates the inadequacy of monolithic, fixed geometry under non-stationary physical and operational environments. This suggests that future designs for both research and field systems should prioritize morphologically adaptive hardware, calibrated and regulated via sensor-driven closed loops, to achieve robust performance across a wide spectrum of scientific and engineering tasks (Oliveira et al., 2022, Griffo et al., 23 May 2026).