Scalable Primitives for Generalized Sensor Fusion in Autonomous Vehicles

TL;DR

This paper introduces Generalized Sensor Fusion, an adaptable neural network architecture for autonomous vehicles that efficiently handles diverse sensor inputs and configurations, enabling scalable deployment across heterogeneous fleets.

Contribution

The paper presents a modular, end-to-end sensor fusion architecture that simplifies experimentation and deployment across different sensor setups and vehicle fleets.

Findings

Near-parity in 3D object detection between high-density LiDAR and low-density LiDAR plus camera setups.

Enables deployment of shared models across heterogeneous vehicle fleets.

Facilitates joint hardware-software design for autonomous vehicle systems.

Abstract

In autonomous driving, there has been an explosion in the use of deep neural networks for perception, prediction and planning tasks. As autonomous vehicles (AVs) move closer to production, multi-modal sensor inputs and heterogeneous vehicle fleets with different sets of sensor platforms are becoming increasingly common in the industry. However, neural network architectures typically target specific sensor platforms and are not robust to changes in input, making the problem of scaling and model deployment particularly difficult. Furthermore, most players still treat the problem of optimizing software and hardware as entirely independent problems. We propose a new end to end architecture, Generalized Sensor Fusion (GSF), which is designed in such a way that both sensor inputs and target tasks are modular and modifiable. This enables AV system designers to easily experiment with different sensor configurations and methods and opens up the ability to deploy on heterogeneous fleets using the same models that are shared across a large engineering organization. Using this system, we report experimental results where we demonstrate near-parity of an expensive high-density (HD) LiDAR sensor with a cheap low-density (LD) LiDAR plus camera setup in the 3D object detection task. This paves the way for the industry to jointly design hardware and software architectures as well as large fleets with heterogeneous configurations.