From MOOS-DAWG'19

Talk: 14-Novitzky

(:notitlegroup:)

Prev-Talk | Next-Talk | All-Talks | Talks-Sorted | MOOS-DAWG'19

Talk-14: pLearn: Behavior Learning with Tensor Flow, Keras, and MOOS-IvP

Michael Novitzky, MIT

The goal of this work is to introduce behavior learning in the MOOS-IvP ecosystem. This work was initially started to support learning behaviors in the complex domain of capture the flag for our Project Aquaticus Testbed. Manually authoring and testing behaviors for autonomous marine vehicles can become tedious and impractical when faced with complex or rapidly changing adversarial situations. We address this problem by learning autonomous behaviors using deep reinforcement learning. We apply deep reinforcement learning, an approach that learns behaviors without relying on explicit vehicle models, to a game of capture the flag with multiple competing vehicles. We have previously presented our integration of deep reinforcement learning with MOOS-IvP, a software suite for marine robotics communication, control, and simulation, that allows the development and execution of behaviors for both underwater and surface vehicles.  We extended MOOS-IvP to create and train a neural net to learn autonomous behaviors for reaching the opponent’s flag while avoiding an adversary exhibiting a defense behavior in simulation.  However, the setup requires serious dependencies — Python and integration with c++ and pHelmIvP….which lead to difficulties on making it portable for other students and colleagues to leverage. We will present our latest integration of pLearn in a Docker container along with simplified installation instructions for greater portability for the research community.

Categories:

Retrieved from https://oceanai.mit.edu/moos-dawg/pmwiki/pmwiki.php?n=Talk.14-Novitzky
Page last modified on June 23, 2019, at 05:01 PM