Eridanus Educational Robotics Program

New Feature of RoboPro Coding!

< Neural Network >

What is a Neural Network in ROBO Pro?

A neural network is a way to make a robot learn from examples instead of following only fixed rules like if–else commands.
In ROBO Pro (the graphical programming software by fischertechnik), a neural network is used to help robots:

-Recognize patterns
-Make decisions
-Improve behavior based on experience

Just like a human brain, the robot uses inputs, processes them, and produces outputs.

How it works (simple idea)

A neural network in ROBO Pro has three main parts:

1. Inputs
These are values from sensors, such as:

Light sensor
Distance sensor
Color sensor
Sound sensor

Example:

Light sensor = bright
Distance sensor = close

2. Processing (the network itself)

The neural network contains neurons
Each neuron has a weight (importance)
ROBO Pro adjusts these weights during training

This is where the “learning” happens.

3. Outputs
These control robot actions, such as:

Motor speed
Turn left or right
Stop or move forward

Example:

Output = turn right
Output = slow down

Training a Neural Network in ROBO Pro

ROBO Pro usually uses supervised learning, which means:

You give input examples
You tell the robot the correct output
The robot adjusts its weights
After repeated training, the robot starts choosing correctly by itself

Example training task:

Input: Light sensor value
Correct output: Follow the black line

Over time, the robot learns line‑following without hard-coded rules.

ROBO Pro does not require math knowledge to use neural networks.

You use graphical blocks, connect sensors and motors, and let the software handle the calculations.

Learn Neural Network in ROBO Pro in our Robotics Summer Camp this May 2026!