Prediction of Autonomous Objects Position Using Neural Networks

The problem of using neural networks for building a vehicle control system is considered in this article. The system includes constructing a safe motion path. The possibility of obtaining information received from global and local sensors of the object as input data is considered. This system is proposed for forecasting position of autonomous object taking into account mutual position relative to other participants of movement. Neural network is proposed to minimize functionality characterizing motion costs and its safety taking into account reconstruction.