SolarSim Simulator — Help & Documentation

Overview

The SolarSim Interactive Simulator is a tool for visualizing and analyzing power harvesting from wearable solar cells under controlled lighting conditions. It provides radiometric simulation with two physics-based models (Orientation Aware and Parallel Disk), configurable light source position and diameter with real-time visualization, and export functionality for simulated power traces in CSV format.

The fundamental power equation:

P = κ \cdot F (a, H) + ϵ

where $P$ is harvested power, $κ$ is the scaling factor, $F$ is the view factor function, and $ϵ$ is noise.

View factor $F$ : The fraction of radiant energy leaving the light source surface that strikes the solar cell surface directly. It depends purely on the geometry—the relative position, orientation, and size of the two surfaces.

For background on view factor theory, see Modest [1] and Bergman et al. [2].

Current assumptions: single dominant emitter, uniform irradiance over cell, and cell positioned below the lamp ( $H > 0$ ).

Quick Start

Everything in the simulator connects to one equation. Hover over each symbol to see what it controls:

P power = κ scale · ℱ model ( a, H geometry ) + ε noise

Power Output

The simulated power curve (purple) in the chart — this is the prediction target.

κ (Kappa)

Scaling factor combining cell efficiency η, cell area A, and light intensity I₀.

View Factor

Choose Orientation Aware or Parallel Disk model.

Geometry

Lateral offset (a) and vertical separation (H) from trajectory and light position.

Power Output

SolarSim Simulator Interface — The SolarSim Simulator interface showing the controls panel (left), 3D visualization (center), and power chart (bottom).

Controls Panel

Control	Description
Trajectory	Select from preloaded trajectorys or choose "Custom" to draw your own
Upload Motion Capture	Load a CSV file with motion capture data (see Input Trajectory Format)
Radiometric Model	Choose between Orientation Aware and Parallel Disk models
Light Position (X, Y, Z)	Adjust the 3D position of the light source in millimeters
Animation Controls	Play, Pause, and Reset the trajectory animation
Speed	Adjust animation playback speed (0.1x to 10x)

3D View

Orbit: Click and drag to rotate the view
Zoom: Scroll wheel to zoom in/out
Pan: Right-click and drag to pan
Trajectory color: Indicates relative power (dark = low, bright = high)

Current State Panel

With more than one emitter in the scene, $a$ , $H$ and $h e t a_{c}$ are reported relative to the primary emitter — the first one that is enabled. Power is always the total across every enabled emitter, so it reflects the whole scene rather than the primary source alone.

Variable	Description
$a$	Lateral displacement from light source axis (mm)
$H$	Vertical distance from light source (mm)
$θ_{c}$	Cell tilt angle relative to light direction (degrees)
Power	Simulated instantaneous power (µW)

View Factor ℱ

The simulator supports two radiometric models for computing the view factor between the light source and the solar cell.

Comparison of Parallel Disk and Orientation Aware models — Parallel Disk assumes cell normals $n_{c}$ always point upward (parallel to emitter normal $n_{e}$ ), while Orientation Aware accounts for actual cell tilt along the trajectory.

Parallel Disk Model

This simplified model assumes the solar cell is always parallel to the light source disk (i.e., facing directly upward). It ignores cell tilt.

Parallel View Factor

F ˉ = H^{2} \int_{0}^{D /2} \int_{0}^{2 π} \frac{r ^{'}}{( r ^{'2} - 2 a r ^{'} cos ϕ + a ^{2} + H ^{2} ) ^{2}} d ϕ d r^{'}

where $H$ is the vertical separation (cell below lamp, $H > 0$ ), $a$ is the lateral offset, $D$ is the disk diameter, and $(r^{'}, ϕ)$ are polar coordinates on the disk.

Best for: Overhead lighting with minimal hand tilt. Requires: Position data only (no cell normals needed).

Orientation Aware Model

This model accounts for the tilt angle of the solar cell relative to the light source. It uses the full radiometric kernel that considers both the emitter and receiver orientations.

Orientation Aware View Factor (approximate)

F = F ˉ \cdot cos θ_{c}

where $F ˉ$ is the parallel disk view factor (Eq. A.5) and $θ_{c}$ is the angle between the cell normal and the direction to the light.

Orientation Aware View Factor (exact)

F = \int_{0}^{D /2} \int_{0}^{2 π} \frac{( R \cdot n _{e} ) ( - R \cdot n _{c} )}{∥ R ∥ ^{4}} r^{'} d ϕ d r^{'}

Full disk integration over the emitter disk of diameter $D$ , where $R = x - x^{'}$ is the vector from emitter patch $x^{'}$ to cell $x$ , $n_{e}$ is the emitter normal, and $n_{c}$ is the cell normal.

Best for: Tilted or rotating cells where orientation matters. Requires: Position + cell normal vectors.

To switch between approximate and exact modes, see Advanced Settings.

Power curve comparison between Orientation Aware and Parallel Disk models — Triangle trajectory: Orientation Aware (left) captures tilt-induced variations with higher R², while Parallel Disk (right) produces a smoother curve that misses these details.

Scale κ

The scaling factor $κ$ converts the dimensionless view factor $F$ into actual power values. It encapsulates all the physical constants and calibration factors of the system.

κ Definition

κ = η \cdot A_{cell} \cdot I_{0}

where $η$ is the solar cell efficiency, $A_{cell}$ is the cell area, and $I_{0}$ is the effective irradiance from the light source.

In a multi-emitter scene each source carries its own $κ_{i}$ , because two lamps may differ in intensity or colour temperature while sharing the same geometry. This simulator fits a single $κ$ to the total view factor when scaling against a reference trace, which is the right choice when the sources share a spectrum and the quantity of interest is trace shape. Fitting per-emitter $κ_{i}$ by non-negative least squares is what the pysolarsim package does when calibrating a real environment.

Configuration Options

Manual Input

Enter a custom $κ$ value directly. Useful when you have a known calibration value from prior experiments or want to explore how scaling affects the output.

Estimate from Physical Parameters

Estimate $κ$ from physical properties when actual values are known:

Solar cell efficiency (typical: 10-25%)
Cell area in cm²
Light source power in Watts

Fit from Real Data

When measured power data is uploaded alongside trajectory data, the simulator can automatically fit $κ$ using least-squares regression:

κ_{opt} = ar g κ min i \sum (P_{real, i} - κ \cdot F_{i})^{2}

where $i$ indexes over trajectory samples, $P_{real, i}$ is the measured power at sample $i$ , and $F_{i}$ is the computed view factor at that position.

Why this option? This is useful for validating the model against experimental data or calibrating $κ$ for a specific hardware setup. It requires real power measurements, which are not always available in simulation-only workflows.

Advanced Settings

Access advanced settings by clicking the gear icon in the Controls panel header.

Simulation Method (Orientation Aware)

Method	Description
Fast (Approximate)	Point source approximation with $cos θ_{c}$ correction
Exact	Full disk integration using Simpson's rule

Light Source Properties

The simulator supports up to four emitters in one scene. Use Add Emitter in the emitter panel to introduce another source; each row carries its own geometry, position, orientation and enable toggle, and each is drawn in a distinct colour in the 3D view.

Shape — Circular (desk lamps, downlights, spotlights) or Rectangular (LED panels, troffers, light strips). Circular emitters are sized by diameter D; rectangular ones by width W and length L, all in millimetres.
Position — X, Y and Z of the emitter centre, in the motion-capture coordinate frame.
Tilt / Azimuth — orientation of the emitter face. Tilt 0° points straight down; azimuth rotates the tilt direction.
Enable toggle — keeps an emitter in the scene but out of the sum, which is useful for isolating one source's contribution.

Total power is the sum of the per-emitter contributions: irradiance from independent sources adds linearly at the cell, so adding an emitter never changes another one's contribution. There is no inter-emitter occlusion or interreflection in the model. The exported CSV records each emitter's shape, position, dimensions and orientation in its header.

Diameter (mm): The diameter of the light source disk. Default is 100mm. Larger values spread the light over a wider area.

Numerical Integration (Exact Mode Only)

These settings control the Simpson's rule grid for disk integration:

Radial divisions ( $n_{r}$ ): Number of divisions along the radius (default: 20)
Angular divisions ( $n_{ϕ}$ ): Number of divisions around the circumference (default: 40)

Higher values increase accuracy but slow down computation. Total integration points = $n_{r} \times n_{ϕ}$ .

Visualization

Color threshold (samples): For trajectories with more samples than this threshold, the drawing animation uses a uniform color instead of power-coded colors for better performance. Default: 4000.

Input Trajectory Format

The simulator accepts CSV files with motion capture data. Several column naming conventions are supported.

Supported Formats

Format 1: Leap Motion Raw

Direct output from Leap Motion SDK:

                            time, leap_palm_x, leap_palm_y, leap_palm_z, leap_palm_normal_x, leap_palm_normal_y, leap_palm_normal_z, v_dc, i_dc
                        

Format 2: Simplified Leap

Shortened column names:

                            timestamp, palm_pos_x, palm_pos_y, palm_pos_z, palm_normal_x, palm_normal_y, palm_normal_z, power
                        

Format 3: Generic XYZ

Minimal format with positions and normals:

x, y, z, nx, ny, nz, power, time

Column Descriptions

Column	Required	Description
`x, y, z` (or variants)	Yes	Palm/cell position in millimeters
`nx, ny, nz` (or variants)	Optional	Cell normal vector (unit vector). Required for Orientation Aware model.
`v_dc`	Optional	Measured voltage (V)
`i_dc`	Optional	Measured current (A). Power = v_dc × i_dc
`power`	Optional	Direct power measurement (alternative to v_dc/i_dc)
`time` or `timestamp`	Optional	Time stamp for each sample

Example CSV

time,leap_palm_x,leap_palm_y,leap_palm_z,leap_palm_normal_x,leap_palm_normal_y,leap_palm_normal_z,v_dc,i_dc
0.00,-45.2,655.8,120.4,0.15,0.92,-0.36,0.48,0.00052
0.01,-44.8,655.1,121.3,0.14,0.93,-0.35,0.49,0.00051
0.02,-43.9,654.2,122.1,0.13,0.94,-0.33,0.47,0.00053
...

Notes

All position values should be in millimeters
Normal vectors should be unit vectors (length ≈ 1)
The cell normal points away from the palm (opposite to palm normal)
Files are processed entirely in your browser — no data is uploaded to servers

Export Format

The file begins with comment lines recording the scene that produced it: one line per emitter giving its shape, position, dimensions, orientation and enabled state, so an export can be traced back to an exact configuration.

When you export simulation results, the CSV contains:

timestamp, simulated_power_W, a_mm, H_mm, theta_deg

Column	Description
`timestamp`	Time in seconds
`simulated_power_W`	Simulated power output in Watts
`a_mm`	Lateral offset from light source (mm)
`H_mm`	Vertical separation from light source (mm)
`theta_deg`	Cell tilt angle in degrees (Orientation Aware model only)

The export file also includes metadata comments at the top with trajectory name, model type, light position, and κ value if applicable.

Citation

If you find this work useful, please cite:

@inproceedings{ghalwash2026solarsim,
  title     = {SolarSim: Physics-Based Simulation of Wearable Solar Harvester Signals for Motion Sensing},
  author    = {Yasien Ghalwash and Abdelwahed Khamis and Sara Khalifa and Moid Sandhu and Raja Jurdak},
  booktitle = {Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies},
  year      = {2026}
}

@inproceedings{ghalwash2026solartrack,
  title     = {SolarTrack: Exploring the Continuous Tracking Capabilities of Wearable Solar Harvesters},
  author    = {Yasien Ghalwash and Abdelwahed Khamis and Moid Sandhu and Sara Khalifa and Raja Jurdak},
  booktitle = {Proceedings of the 24th IEEE International Conference on Pervasive Computing and Communications (PerCom)},
  year      = {2026}
}

References

Modest, M. F., & Mazumder, S. (2021). Radiative Heat Transfer. Academic Press.
Bergman, T. L., Lavine, A. S., Incropera, F. P., & DeWitt, D. P. (2017). Fundamentals of Heat and Mass Transfer (8th ed.). Wiley. — Chapter 13: Radiation Exchange Between Surfaces.

Contact

For questions or feedback, open an issue on GitHub or email abdelwahed.khamis@data61.csiro.au.