Light Signal Decoding Using Low-Complexity ANN Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Li-Fi technologies require high computational power for data transmission, hindering widespread adoption due to resource constraints in everyday devices.
Innovation Solution
A light signal decoding device and method using a light signal encoding device with a memory and processor to generate multi-wavelength light signals, combined with an artificial neural network (ANN) model to decode these signals efficiently, reducing computational demands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If Li-Fi technology is used for secure and fast data transmission, then data transmission speed and security are improved, but computational power requirements increase
Solution Approach 1:
The patent changes the parameter of computational complexity by using a simplified neural network architecture with only one hidden layer containing 8 nodes, reducing the computational power requirements while maintaining fast data transmission speeds through efficient light signal encoding and decoding
Solution Approach 2:
The patent creates a pre-trained neural network model that can be copied and deployed on devices with lower processing power, allowing the benefits of Li-Fi to be replicated without requiring high computational resources at runtime
2Measurement precision
If neural network models are used to decode light signals, then data decoding accuracy is improved, but device complexity increases
Solution Approach 1:
The patent extracts only the essential components needed for accurate decoding by using a minimal neural network architecture with one hidden layer and 8 nodes, removing unnecessary complexity while maintaining decoding accuracy
Solution Approach 2:
The patent segments the decoding process into a pre-trained neural network model that handles complex pattern recognition, while the device itself only needs to capture light signals and provide minimal input data, dividing the complexity between the model and the device
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables fast and efficient data transmission with reduced computational requirements, allowing Li-Fi technology to be utilized in devices with lower processing power.
Implementation Method 1
a light signal generating device configured to generate at least a two different wavelength light signal
Implementation Method 2
an image sensor configured to capture at least two different wavelength light signal
Data Source
AI summary
Invention relates to a light signal decoding device and a light signal decoding method. The light signal decoding device comprises an image sensor configured to capture at least two different wavelength light signal, a memory configured to store an ANN (Artificial Neural Network) model. The ANN model comprises at least three input neurons, only two hidden layers and at least two output neurons as at least two data bits. The device comprises a processor configured to transform the captured at least two different wavelength light signal into at least two numeric values, sum two numeric values to obtain a C (Clear colour) value, and provide each numeric value and the C value as input neurons to the ANN model, wherein the ANN model apply Leaky ReLU activation function and ReLU activation function to obtain at least two output neurons as at least two data bits.