The Autonomous Object Trajectory Correction Using Neural Networks

The issue of using neural networks to construct and correct the transport object trajectory is considered in this article. The control system, which includes the analysis of general unchangeable information received from global sensors, is considered. The decision on the need to change the trajectory is made depending on the data received from sensors located directly on the object. A method is proposed for correcting the autonomous object route depending on the occurrence of stationary obstacles during movement. A neural network is defined that minimizes the functional characterizing the cost of movement and its safety, taking into account the rebuilding.