Convolutional Neural Networks Training for Autonomous Robotics

The article discusses methods for accelerating the operation of convolutional neural networks for autonomous robotics learning. The analysis of the theoretical possibility of modifying the neural network learning mechanism is carried out. Classic semiotic analysis and the theory of neural networks is proposed to union. An assumption is made about the possibility of using the symmetry mechanism to accelerate the training of convolutional neural networks. A multilayer neural network to represent how space is an attempt has been made. The conclusion was based on the laws on the plane obtained earlier. The derivation of formulas turned out to be impossible due to the problems of modern mathematics. A new approach is proposed, which involves combining the gradient descent algorithm and the stochastic completion of convolutional filters by the principles of symmetries. The identified algorithms allow increasing the learning rate from 5% to 15%, depending on the problem that the neural network solves.

Micromouse 3D simulator with dynamics capability: a Unity environment approach

The micromouse competition has been gaining prominence in the robotic atmosphere, due to the challenging and multidisciplinary characteristics provided by the teams’ duels, being a gateway for those who intend to deepen their studies in autonomous robotics. In this context, this paper presents a realistic micromouse simulator developed with Unity software, a widely game engine with dynamics and 3D development platform used. The developed simulator has hardware-in-the-loop capabilities, aims to be simple to use, it can be customizable, and designed to be as similar as possible to the real robot configurations. In this way, the proposed simulator requires few modifications to port the microcontroller code to a real robot. Therefore, the framework presented in this work allows the user to simulate the development of new algorithm strategies dedicated to competition and also hardware updates. The simulation supports several mazes, from previous competitions and has the possibility to add different mazes elaborated by the user. Thus, the features and functionality of the simulator can serve to accelerate the project’s development of the beginning and advanced competitors, using real models to reduce the gap between the mouse robot behavior in the simulation and the reality. The developed simulation environment is available to the community.