Autonomous vehicles are a key element of the automotive industry, where the impact of the human factor on the condition of the vehicle and driving is minimized. An important element is the analysis of vehicular condition, which allows maintainence of its value and correct operation. We propose a system based on the analysis of the image of vehicles, which determines whether there is any damage. For this purpose, we propose a new model of a Convolutional Neural Network (CNN) that has 0. 395M trained values. The architecture of the network is adapted to the analysis of spatial features that allow networks to be adapted to analyze primarily vehicular shape and orientation in relation to other objects. The model also implements spatial dropout and regularization techniques for preventing overtraining and maintaining model generalization. The modeled architecture contributes to obtaining high classification accuracy at 94.78% using a public database and exceeding metrics of known transfer learning models.

The recent development of Autonomous Guided Vehicles (AGV) use in industry has resulted in the need to model new solutions based on the latest technological achievements. One of the areas worth attention and development is Quality of Service (QoS) in relation to communication between vehicles. QoS makes it possible to divide the bandwidth in such a way that tasks performed by devices are completed with a certain priority. However, in order to manage these resources effectively, it is necessary to anticipate available network throughput. Therefore, this paper presents a neural-based model to ensure throughput prediction for AGV. The proposed solution assumes the use of information on both historical throughput values and data obtained from other sensors that AGV are equipped with. Therefore, the idea is to integrate two neural networks with another network, which is supposed to predict the result based on these two previously obtained predictions. Ultimately, prediction results with a Root Mean Squared Error (RMSE) of 0.1 for the downlink and 1.6 for the uplink were obtained.

Automated guided vehicles (AGV) allow for the automation of operations in warehouse environments. From an application point of view, vehicles can use sensors to move and perform a variety of tasks, including moving objects. In this paper, we focus on the analysis of the environment and the preparation of data for vehicle navigation. The proposed solution is based on two paths of action. In the first, the image of the room is processed by heuristics to locate the robot’s target - the palette. The found pattern allows one to locate the destination as well as create a mask. The mask can be used to train the U-Net network. When a network is trained, the use of heuristics for pallet location can be omitted. Locating the target allows the image to be processed to obtain an AGV navigation map. The proposed solution based on heuristics and U-Net networks has been described and tested in simulations to indicate the potential of the proposed approach.