Neural network-based optimization of intelligent odometry system in an autonomous robot for accurate position and orientation estimation

Abstract

The purpose of this study is to determine the accurate orientation and position of the intelligent dummy system for the Verdino autonomous robot. The position and orientation of the autonomous robot are calculated using an odometric mathematics model derived from the robot's trajectory equations. The odometry encoders are the system inputs, and the wheel diameter and distance are used to parameterize the model. This model is used to get the optimum nominal parameters by doing a minimal square minimization. This model is updated with a measurement of the real-time wheel diameter to improve the accuracy of the results. Based on data, an odometric model is trained using a neural network model. This neural network is put to the test, and the results indicate that by employing a smart odometry system, the neural network can improve estimate accuracy.