Simulation analysis of a highway dnn for autonomous forklift dispatching

With the proliferation of autonomous vehicles in warehouse applications, there are several challenges that face researchers including precision indoor localization, navigation, obstacle avoidance, path planning, and task selection decisions. This paper addresses the issue of task selection decision. Specifically, we develop a deep learning methodology for task selection for fleet of autonomous vehicles in a warehouse environment. The autonomous vehicles select a task from a list of tasks, considering current vehicle traffic, potential travel paths, and the task potential task locations. We implement a highway deep neural network (DNN) for the task selection process. To evaluate the methodology, we conducted a simulation-based experiment to generate various scenarios and test the capabilities of the DNN. The results of the simulation-based experiment show that our deep learning method performs well under the given conditions.