Towards Generalization in Target-Driven Visual Navigation by Using Deep Reinforcement Learning
| Autor: | Giacomo Mezzetti, Paolo Valigi, Gabriele Costante, Mario Luca Fravolini, Alessandro Devo |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
0209 industrial biotechnology
Computer science business.industry Deep Learning in Robotics and Automation Robotics 02 engineering and technology Simultaneous localization and mapping Computer Science Applications Task (project management) Visualization 020901 industrial engineering & automation Target-Driven Visual Navigation Control and Systems Engineering Human–computer interaction Visual-Based Navigation Generalization (learning) Task analysis Reinforcement learning Artificial intelligence Target-Driven Visual Navigation Visual-Based Navigation Deep Learning in Robotics and Automation Visual Learning Electrical and Electronic Engineering business Visual learning Visual Learning |
| Zdroj: | IEEE Transactions on Robotics. 36:1546-1561 |
| ISSN: | 1941-0468 1552-3098 |
| DOI: | 10.1109/tro.2020.2994002 |
| Popis: | Among the main challenges in robotics, target-driven visual navigation has gained increasing interest in recent years. In this task, an agent has to navigate in an environment to reach a user specified target, only through vision. Recent fruitful approaches rely on deep reinforcement learning, which has proven to be an effective framework to learn navigation policies. However, current state-of-the-art methods require to retrain, or at least fine-tune, the model for every new environment and object. In real scenarios, this operation can be extremely challenging or even dangerous. For these reasons, we address generalization in target-driven visual navigation by proposing a novel architecture composed of two networks, both exclusively trained in simulation. The first one has the objective of exploring the environment, while the other one of locating the target. They are specifically designed to work together, while separately trained to help generalization. In this article, we test our agent in both simulated and real scenarios, and validate its generalization capabilities through extensive experiments with previously unseen goals and unknown mazes, even much larger than the ones used for training. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|