Vision-Based Autonomous Drone Control using Supervised Learning in Simulation
This addresses challenges for autonomous drones in resource-limited settings, but it is incremental as it builds on existing supervised learning techniques with simulation data.
The paper tackles autonomous navigation and landing for micro aerial vehicles in indoor, GPS-denied environments by proposing a vision-based control approach using supervised learning with a CNN trained on simulation data, achieving successful navigation toward a landing platform in simulations with shorter training times than reinforcement learning methods.
Limited power and computational resources, absence of high-end sensor equipment and GPS-denied environments are challenges faced by autonomous micro areal vehicles (MAVs). We address these challenges in the context of autonomous navigation and landing of MAVs in indoor environments and propose a vision-based control approach using Supervised Learning. To achieve this, we collected data samples in a simulation environment which were labelled according to the optimal control command determined by a path planning algorithm. Based on these data samples, we trained a Convolutional Neural Network (CNN) that maps low resolution image and sensor input to high-level control commands. We have observed promising results in both obstructed and non-obstructed simulation environments, showing that our model is capable of successfully navigating a MAV towards a landing platform. Our approach requires shorter training times than similar Reinforcement Learning approaches and can potentially overcome the limitations of manual data collection faced by comparable Supervised Learning approaches.