Learning Continuous Control Policies for Information-Theoretic Active Perception

TL;DR

This paper introduces a novel method for learning continuous control policies for active perception tasks like landmark localization and exploration, leveraging an information-theoretic cost and neural networks.

Contribution

It presents a unified approach combining mutual information maximization, Kalman filtering, and attention-based neural networks for active perception and exploration.

Findings

Effective in simulated landmark localization tasks

Outperforms benchmark methods in exploration and localization

Integrates active mapping with control policy learning

Abstract

This paper proposes a method for learning continuous control policies for active landmark localization and exploration using an information-theoretic cost. We consider a mobile robot detecting landmarks within a limited sensing range, and tackle the problem of learning a control policy that maximizes the mutual information between the landmark states and the sensor observations. We employ a Kalman filter to convert the partially observable problem in the landmark state to Markov decision process (MDP), a differentiable field of view to shape the reward, and an attention-based neural network to represent the control policy. The approach is further unified with active volumetric mapping to promote exploration in addition to landmark localization. The performance is demonstrated in several simulated landmark localization tasks in comparison with benchmark methods.