Category EGCH P07 Determining light-source location using Machine Learning and

Solar Cells


The current generated by a single solar cell is highly sensitive to the

angle, the intensity, and the color of the light incident upon it. In this

project, I want to investigate whether this property can be utilized to

design an “intelligent” solar panel that can “predict” the position of a

light source incident upon it (and, possibly, also its distance and its

color). A potential application of such an “intelligent” panel would be as

a solar tracker, facilitating the determination of the position of the Sun

and feeding positional data to the tilting control unit.


My hypothesis is that if an array of solar cells can be properly

designed, then the electrical response of each cell in the array will be

different for different locations of an incandescent source of light

pointed towards it. These electrical response values can then be read

by a computer using proper sensors. The latter data, along with the

positional information of the light source can then be used to train a

supervised Machine Learning algorithm. The generated model can then

be used to predict the position of any incandescent light source

pointed at the array.


1.Use Ruthenizer-535 synthetic dye to create dye-sensitized solar cells

(DSSCs). I choose to use DSSCs because I can produce them at home.

Moreover, DSSCs are environmentally friendly, cheaper to

manufacture than their silicon-based counterparts, and work in low-light


2.Use 6-9 such DSSCs to create an array so that each cell responds

differently to an incandescent light source pointed at the array.

3.Connect each cell to the ADC input of an Arduino.

4.Read the voltage-drop on a 560-ohm resistor connected to each cell

using an Arduino program.

5.Record the position of the light source in a 2-dimensional / 3-

dimensional unit.

6.Use MatLab programs to train a Machine Learning algorithm and

generate a model for light-source-position prediction.

First Previous Next Last