Comparative Analysis of Neural Network Models for Error Probability Prediction in Vehicular Communication

Reliable data transfer is a tedious task for various applications that involve users in motion. In vehicular communication systems, for reliable and seamless data transfer an accurate estimation strategy is needed. The principal objective involves construction of the predictive models that can estimate error probabilities on parameters like signal strengths, timestamps, and sender identification outcomes. To establish a proper strategy for vehicle scenarios while maintaining the system’s effectiveness, we present an analysis of our system with error probabilities, exploring the relationship between signal strength, neural network (NN) architectures, and error prediction accuracy. Data generated is used to simulate communication messages. Furthermore, this paper boards to calculate the effectiveness of different NN architectures, including single and multi-layer perceptrons, interpreting key relations embedded in data and providing precise error probability predictions. Through a series of insights, we observed signal strength and NN impact on error predictions. Our analysis improves the complex error pattern occurrences and identifies the most suitable architecture for accurate predictions.