Backprop KF: Learning Discriminative Deterministic State Estimators

TL;DR

This paper introduces Backprop KF, a discriminative, deterministic state estimator trained via gradient descent, capable of processing complex sensory data like raw images, outperforming traditional probabilistic filters and standard RNNs.

Contribution

It proposes a novel training method for discriminative state estimators using a deterministic computation graph, enabling effective learning from rich sensory inputs.

Findings

Significant performance improvements over generative filters.

Outperforms standard recurrent neural networks on visual odometry.

Effective training on raw camera images with CNNs.

Abstract

Generative state estimators based on probabilistic filters and smoothers are one of the most popular classes of state estimators for robots and autonomous vehicles. However, generative models have limited capacity to handle rich sensory observations, such as camera images, since they must model the entire distribution over sensor readings. Discriminative models do not suffer from this limitation, but are typically more complex to train as latent variable models for state estimation. We present an alternative approach where the parameters of the latent state distribution are directly optimized as a deterministic computation graph, resulting in a simple and effective gradient descent algorithm for training discriminative state estimators. We show that this procedure can be used to train state estimators that use complex input, such as raw camera images, which must be processed using expressive nonlinear function approximators such as convolutional neural networks. Our model can be viewed as a type of recurrent neural network, and the connection to probabilistic filtering allows us to design a network architecture that is particularly well suited for state estimation. We evaluate our approach on synthetic tracking task with raw image inputs and on the visual odometry task in the KITTI dataset. The results show significant improvement over both standard generative approaches and regular recurrent neural networks.