Deep Reinforcement Learning for Goal-directed Visual Navigation

Deep Reinforcement Learning for Goal-directed Visual Navigation

Kisantal, Máté (TU Delft Aerospace Engineering; TU Delft Control & Operations)

de Croon, G.C.H.E. (mentor)
van Hecke, K.G. (mentor)

Delft University of Technology

Aerospace Engineering | Control & Simulation

2018-02-23

Safe navigation in a cluttered environment is a key capability for the autonomous operation of Micro Aerial Vehicles (MAVs). This work explores a (deep) Reinforcement Learning (RL) based approach for monocular vision based obstacle avoidance and goal directed navigation for MAVs in cluttered environments. We investigated this problem in the context of forest flight under the tree canopy.

Our focus was on training an effective and practical neural control module, that is easy to integrate into conventional control hierarchies and can extend the capabilities of existing autopilot software stacks. This module has the potential to greatly improve the autonomous capabilities of MAVs, and their applicability for many interesting real world use-cases. We demonstrated training this module in a visually highly realistic virtual forest environment, created with a state-of-the-art computer game engine.

reinforcement learning
deep reinforcement learning
artificial intelligence
machine learning
computer vision
MAV
UAV
MAVLAB
drone
autonomous navigation
Autonomous Vehicles
deep learning
neural networks

http://resolver.tudelft.nl/uuid:07bc64ba-42e3-4aa7-ba9b-ac0ac4e0e7a1

Student theses

master thesis

© 2018 Máté Kisantal