Integrating Algorithmic Planning and Deep Learning for Partially Observable Navigation

Abstract

We propose to take a novel approach to robot system design where each building block of a larger system is represented as a differentiable program, i.e. a deep neural network. This representation allows for integrating algorithmic planning and deep learning in a principled manner, and thus combine the benefits of model-free and model-based methods. We apply the proposed approach to a challenging partially observable robot navigation task. The robot must navigate to a goal in a previously unseen 3-D environment without knowing its initial location, and instead relying on a 2-D floor map and visual observations from an onboard camera. We introduce the Navigation Networks (NavNets) that encode state estimation, planning and acting in a single, end-to-end trainable recurrent neural network. In preliminary simulation experiments we successfully trained navigation networks to solve the challenging partially observable navigation task.

arXiv commentMLPC workshop, ICRA 2018

References in corpus (10)

• Deep Recurrent Q-Learning for Partially Observable MDPs
• DESPOT: Online POMDP Planning with Regularization
• OptNet: Differentiable Optimization as a Layer in Neural Networks
• Task-based End-to-end Model Learning in Stochastic Optimization
• Backprop KF: Learning Discriminative Deterministic State Estimators
• Value Prediction Network
• Path Integral Networks: End-to-End Differentiable Optimal Control
• Learning to Search with MCTSnets
• TreeQN and ATreeC: Differentiable Tree-Structured Models for Deep Reinforcement Learning
• Particle Filter Networks with Application to Visual Localization

Cited by in corpus (4)

• Neural Predictive Belief Representations
• TASKOGRAPHY: Evaluating robot task planning over large 3D scene graphs
• DualSMC: Tunneling Differentiable Filtering and Planning under Continuous POMDPs
• When is Particle Filtering Efficient for Planning in Partially Observed Linear Dynamical Systems?