Active Classification of Moving Targets with Learned Control Policies

TL;DR

This paper introduces a reinforcement learning-based attention architecture for drones to actively classify multiple moving targets by selecting informative viewpoints, effectively handling occlusions and target dynamics.

Contribution

The paper presents a novel RL-trained attention model that predicts optimal viewpoints for drone-based classification of moving targets, addressing the challenge of black-box classifier integration.

Findings

Outperforms baseline methods in target classification accuracy

Generalizes to unseen scenarios and target movement patterns

Scales effectively with increasing number of targets

Abstract

In this paper, we consider the problem where a drone has to collect semantic information to classify multiple moving targets. In particular, we address the challenge of computing control inputs that move the drone to informative viewpoints, position and orientation, when the information is extracted using a "black-box" classifier, e.g., a deep learning neural network. These algorithms typically lack of analytical relationships between the viewpoints and their associated outputs, preventing their use in information-gathering schemes. To fill this gap, we propose a novel attention-based architecture, trained via Reinforcement Learning (RL), that outputs the next viewpoint for the drone favoring the acquisition of evidence from as many unclassified targets as possible while reasoning about their movement, orientation, and occlusions. Then, we use a low-level MPC controller to move the drone to the desired viewpoint taking into account its actual dynamics. We show that our approach not only outperforms a variety of baselines but also generalizes to scenarios unseen during training. Additionally, we show that the network scales to large numbers of targets and generalizes well to different movement dynamics of the targets.