AUTONOMOUS DRONE FOLLOWING WITH REINFORCEMENT LEARNING
Project developed in the Massachusetts Institute of Technology (MIT).
MITDeep Reinforcement LearningComputer VisionSynthetic Data
Overview
This project aimed at the validation of an end-to-end and vision-based technique for non-cooperative multirotor following. The complete approach used synthetic information in the context of deep and reinforcement learning techniques. The strategy was partially developed in the Massachusetts Institute of Technology (MIT).
My Role
Researcher
Technologies & Skills
Deep Reinforcement LearningSynthetic DataComputer VisionDrone Tracking
Key Achievements
- ✓End-to-end vision-based multirotor following
- ✓Developed at MIT
- ✓Published co-authored research paper
- ✓Novel use of synthetic information for training a starndard object detector