cs.ROOct 27, 2026
SaveIntegrated Design and Reproducible Feasibility Analysis of a Four-Wheel Rover with Packet-Commanded Imaging, LiDAR Ranging, and Learning-Assisted Target Recognition
Organizations: BAF Shaheen College Kurmitola · St. Joseph Higher Secondary School · Dhaka Residential Model College · South Point School & College
Abstract
This study develops an integrated, reproducible design assessment for a four-wheel educational rover intended for the high-school remote-control division of the 2027 NASA Human Exploration Rover Challenge. The architecture combines commercial differential drive control, independently sprung wheel carriers, separate drive and task batteries, and a Raspberry Pi-based subsystem that executes received imaging and ranging commands. The analysis reconciles component lists with dimensioned computer-aided design views, derives mobility and electrical constraints, and connects sensor geometry to command execution. At the estimated mass of 10.919 kg, a 215 mm wheel rotating at 200 RPM produces a theoretical rolling speed of 8.105 km/h; consequently, the nominal motor label does not establish compliance with an 8 km/h design ceiling. A steady 20-degree climb requires at least 1.255 newton-meters per wheel under an assumed rolling-resistance coefficient of 0.10 and equal torque sharing. At 50 ft, an ideal 12-inch feature spans approximately 71 pixels across a full-resolution Camera Module 3 Standard image, while a 1-degree pointing error displaces the ranging ray by approximately 266 mm. An executed synthetic benchmark compares color-geometry rules and two linear support-vector classifiers on 2,400 generated board crops with scene-disjoint evaluation. A separate reference command executor is checked using mocked hardware. These software results and analytical sensitivities support a testable development plan; the study does not report physical rover trials. The principal contribution is a traceable connection between component choices, failure mechanisms, quantitative design limits, and reproducible verification tasks.
Figures & tables
Figure 1: Supplied rover CAD views: assembly and side profile. These are design renderings supplied by the team, not photographs of a fabricated prototype. The latest dimensioned design takes precedence over earlier concept images.
| Quantity | Adopted value or interpretation | Evidence status |
|---|---|---|
| Overall length × width | 655 × 630 mm | Supplied design dimensions |
| Height to camera top | 421 mm, antenna excluded | Supplied design dimension |
| Wheels | Four; 215 mm diameter × 90 mm width | Supplied design dimensions |
| Rim / tire radial thickness | 188 mm / 13.5 mm | Dimension closure verified |
| Estimated fully equipped mass | 10.919 kg | Team-supplied theoretical estimate; unweighed |
| Track / wheelbase | 540 / 440 mm | Conditional inference from outer dimensions |
Table 1: Adopted design quantities and their evidence status.
| Handbook consideration | Design interpretation | Evidence needed before acceptance |
|---|---|---|
| Excursion-ready cube, Rule 5.5.4 | 762 mm on each axis | Measure complete rover, including antenna |
| High-school manufacture, Rule 6.4.2.1 | Team-designed/manufactured chassis, wheels and tires; permitted commercial drive/suspension parts | Manufacturing records and final configuration |
| Primary drive control, Rule 6.4.3 | Commercial transmitter, receiver and ESC control | Demonstration of configured control and failsafe |
| Speed, Rule 6.4.4 | Use 8 km/h as the conservative ceiling associated with the stated 5 mph limit | Full-charge maximum-speed test and limiter verification |
| Mass, Rules 6.4.5 and Section 8.2 | 60 lb absolute limit; 20 lb is a scoring objective | Weigh excursion-ready vehicle |
| Battery protection, Rules 6.4.6–6.4.7 | Separate insulated enclosures; main overcurrent protection | Wiring and enclosure inspection |
Table 2: Handbook requirements mapped to verification needs.
Figure 2: New Blender reconstruction of the proposed rover. Geometry is an approximate reconstruction from the supplied CAD views and dimensions; the rendering illustrates packaging and appearance, not a manufactured vehicle or structural simulation.
Figure 3: Published FDM TPU tire used as a manufacturing comparison. Reproduced from J. Wang et al., original Fig. 12 [ 12 ] , under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is the cited authors’ tire, not a photograph of the present rover wheel; its properties are not transferred to this design.
Figure 4: Editable Canva drive-system block diagram. It separates receiver commands, the single active BEC supply, battery distribution, and reversible motor outputs. Each motor-output connection represents a conductor pair; detailed connector pinouts remain a wiring-drawing task.
Figure 5: Published skid-steer instantaneous-center geometry, reproduced from T. Wang et al., original Fig. 1 [ 3 ] , under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). The symbols and geometry belong to that paper. The figure motivates measured slip-aware kinematics; no fitted coefficient is copied into the present rover model.
Figure 6: Calculated rolling speed versus wheel RPM and an explicitly conditional voltage-scaling scenario. These curves use the nominal 215 mm diameter; they are not measured speed traces.
| Slope | Required total force (N) | Wheel-output torque per wheel (N·m) | Minimum friction coefficient |
|---|---|---|---|
| 0° | 10.708 | 0.288 | 0.100 |
| 10° | 29.139 | 0.783 | 0.276 |
| 15° | 38.057 | 1.023 | 0.368 |
| 20° | 46.685 | 1.255 | 0.464 |
| 25° | 54.958 | 1.477 | 0.566 |
| 30° | 62.813 | 1.688 | 0.677 |
Table 3: Calculated steady straight-climb lower bounds at assumed rolling resistance 0.10.
Figure 7: Analytical straight-climb sensitivities: wheel-output torque, friction-limited grade, and mechanical power. Rolling resistance, equal torque sharing and steady motion are declared assumptions. Turning and transient losses are excluded.
Figure 8: Conditional static tip-angle sensitivity to center-of-mass height. The support rectangle is inferred from the supplied external dimensions; the center-of-mass height is swept rather than measured.
Figure 9: Blender illustration of one wheel module, showing the paired coils and single damper arrangement. The scene reconstructs the visible mechanism; spring rate, damping coefficient, bearing fits and full suspension travel remain to be measured or finalized.
Figure 10: Local suspension sensitivity for two springs and one damper per wheel. Spring rates, motion ratios and sprung masses are illustrative parameter sweeps; none is a measured response of the proposed hardware.
Figure 11: Editable Canva task-power diagram. The separate positive rails and the low-side battery guard are explicit. Branch labels describe proposed hardware, not measured current or validated fuse coordination.
| Task branch | Illustrative output power (W) | Calculated battery-side current (A) | Proposed input fuse (A) |
|---|---|---|---|
| Pi and USB branch | 25.75 | 2.885 | 5 |
| Pan-servo branch | 15.00 | 1.681 | 3 |
| LiDAR branch | 2.00 | 0.224 | 1 |
| Tilt-servo branch | 5.20 | 0.583 | 1 |
| Total | 47.95 | 5.373 | 7.5 main |
Table 4: Illustrative task-branch load and battery-side current budget.
Figure 12: Calculated energy and current scenarios for separate drive and task packs. Assumed usable energy and conversion efficiency are shown; the figures do not represent measured battery discharge or load traces.
| Token | Required operation represented in the executor | State consequence |
|---|---|---|
| AAA | Begin a task sequence | Open a new record after successful initialization |
| B / C | Turn field of view 30° right / left | Update pan only after successful motion |
| D | Capture and save a new photograph | Use current filter; attach capture date/time |
| E / F | Select red / blue filter | Replace the prior filter selection |
| G | Select no filter | Restore normal-color output |
| H | Measure target distance | Record a fresh, valid result in feet |
Table 5: Received task-command semantics and execution state.
Figure 13: Editable Canva task-interface diagram showing the receiving chain, computer, camera, LiDAR, servo control, storage and display. This is an integration specification; exact device revision and pin-level checks remain necessary.
Figure 14: Editable Canva capture pipeline. The command queue, current filter, settled view, fresh frame, timestamp, storage and onboard review are represented separately so that an image can be traced to its command.
| Range | Angular width of a 12-inch feature | Full 4608-pixel width estimate | Same full view scaled to 1920 pixels |
|---|---|---|---|
| 10 ft | 5.725° | 354.78 px | 147.83 px |
| 25 ft | 2.292° | 141.91 px | 59.13 px |
| 50 ft | 1.146° | 70.96 px | 29.57 px |
Table 6: Ideal sampling of a 12-inch feature using the complete camera view.
Figure 15: Blender illustration of the photography task with the rover stationary and illustrative target placements. The scene is a visualization of the intended operation, not a rendered dataset used for the classification experiment and not an observed competition run.
Figure 16: Editable Canva ranging pipeline with target association, calibration, frame validation, fresh sampling, bounded retry and explicit unavailable-result handling. A plausible distance alone is insufficient evidence of a correct target measurement.
Figure 17: Blender ranging illustration. The beam and annotated board layout are schematic visualization elements; distances in the scene are illustrative, not recorded TF03 measurements or a surveyed NASA course.
Figure 18: Calculated camera sampling, nominal beam-width geometry and aiming-error sensitivity. These are deterministic geometric calculations with declared sensor specifications, not calibration residuals from the rover.
Figure 19: Editable Canva command-state and fault-flow diagram. Reception and execution are decoupled, and invalid, out-of-range or failed operations remain visible in the record.
| Split | Scene groups | Images | Purpose |
|---|---|---|---|
| Training | 300 | 1,200 | Fit feature standardization and SVM parameters |
| Validation | 100 | 400 | Select declared hyperparameters |
| Nominal test | 100 | 400 | Evaluate unseen scenes from nominal generator ranges |
| Shifted test | 100 | 400 | Evaluate altered backgrounds and stronger corruptions |
Table 7: Scene-disjoint synthetic dataset allocation.
| Nuisance parameter | Nominal range | Shifted range |
|---|---|---|
| Board width / image width | 0.72–0.88 | 0.64–0.83 |
| Rotation | 30° to +30° | 55° to +55° |
| Per-corner projective jitter | ±0.045 image widths | ±0.10 image widths |
| Exposure multiplier | 0.72–1.20 | 0.38–1.40 |
| Per-channel gain | 0.85–1.15 | 0.58–1.40 |
| Blur standard deviation | 0–0.65 output pixels | 0.65–1.65 output pixels |
Table 8: Selected appearance ranges in the nominal and shifted generators.
Figure 20: Examples from the executed synthetic dataset. Nominal and shifted board crops include target markers and no-target negatives. These are procedurally generated images, not photographs from the rover camera or Blender task scenes.
| Method | Nominal accuracy | Nominal macro-F1 | Shifted accuracy | Shifted macro-F1 |
|---|---|---|---|---|
| HSV + geometry | 95.00% | 0.9512 | 71.75% | 0.7343 |
| HSV histogram + SVM | 82.00% | 0.7892 | 61.75% | 0.5838 |
| HOG + HSV + SVM | 82.00% | 0.7890 | 48.50% | 0.4784 |
Table 9: Executed classification results on the two synthetic test sets.
Figure 21: Executed classification results under nominal and shifted synthetic distributions. Error bars use whole-scene bootstrap resampling. The explicit baseline performs better than both learned models; the graph is not a field-performance comparison.
Figure 22: Confusion matrices from the executed synthetic benchmark. Counts correspond to 100 images per true class in each test set. The no-target column/row behavior is essential when assessing whether classification could assist board association.
| Subsystem | Verification activity | Record needed | Current status |
|---|---|---|---|
| Envelope and mass | Measure fully equipped rover, including antenna | Dimension sheet and scale record | CAD dimensions and mass estimate only |
| Propulsion | Full-charge speed, sustained climb, repeated turns, stop/failsafe | Speed, current, temperature, stop behavior | Analytical scenarios only |
| Suspension | Loaded corner force–travel and articulation checks | Motion ratios, end stops, corner weights | CAD concept and sensitivity model |
| Printed wheels | Coupon, hub, tire/interface and repeated-load tests | Process settings, forces, displacement, damage | Literature-informed test plan |
| Task power | Concurrent camera, display and servo operation | Rail voltage, pack current, temperature | Energy/current scenarios only |
| RF reception | Complete sequence with motors and converters operating | Received packets, errors, queue timestamps | Architecture and adapter contract |
Table 10: Integrated verification activities and present evidence.