Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FuzzSense: Towards A Modular Fuzzing Framework for Autonomous Driving Software

Andrew Roberts, Lorenz Teply, Mert D. Pese, Olaf Maennel, Mohammad Hamad, Sebastian Steinhorst

TL;DR

FuzzSense addresses robustness evaluation for autonomous driving software by enabling ensemble fuzzing across sensors, scenarios, and vehicle dynamics. It introduces a modular framework with components such as the Fuzzing Broker, Orchestrator, Mutator, Scenario and Sensor Fuzzers, and Oracle/Evaluation, and demonstrates a LiDAR fuzzing plug-in within AWSIM and Autoware.Universe. The LiDAR fuzzing experiments reveal that manipulated points can be misinterpreted as obstacles or trigger braking, highlighting vulnerabilities in AD stacks. The work contributes an open-source platform to catalyze community discussion and development of AD-specific fuzzing tools and a shared fuzzing framework.

Abstract

Fuzz testing to find semantic control vulnerabilities is an essential activity to evaluate the robustness of autonomous driving (AD) software. Whilst there is a preponderance of disparate fuzzing tools that target different parts of the test environment, such as the scenario, sensors, and vehicle dynamics, there is a lack of fuzzing strategies that ensemble these fuzzers to enable concurrent fuzzing, utilizing diverse techniques and targets. This research proposes FuzzSense, a modular, black-box, mutation-based fuzzing framework that is architected to ensemble diverse AD fuzzing tools. To validate the utility of FuzzSense, a LiDAR sensor fuzzer was developed as a plug-in, and the fuzzer was implemented in the new AD simulation platform AWSIM and Autoware.Universe AD software platform. The results demonstrated that FuzzSense was able to find vulnerabilities in the new Autoware.Universe software. We contribute to FuzzSense open-source with the aim of initiating a conversation in the community on the design of AD-specific fuzzers and the establishment of a community fuzzing framework to better target the diverse technology base of autonomous vehicles.

FuzzSense: Towards A Modular Fuzzing Framework for Autonomous Driving Software

TL;DR

FuzzSense addresses robustness evaluation for autonomous driving software by enabling ensemble fuzzing across sensors, scenarios, and vehicle dynamics. It introduces a modular framework with components such as the Fuzzing Broker, Orchestrator, Mutator, Scenario and Sensor Fuzzers, and Oracle/Evaluation, and demonstrates a LiDAR fuzzing plug-in within AWSIM and Autoware.Universe. The LiDAR fuzzing experiments reveal that manipulated points can be misinterpreted as obstacles or trigger braking, highlighting vulnerabilities in AD stacks. The work contributes an open-source platform to catalyze community discussion and development of AD-specific fuzzing tools and a shared fuzzing framework.

Abstract

Fuzz testing to find semantic control vulnerabilities is an essential activity to evaluate the robustness of autonomous driving (AD) software. Whilst there is a preponderance of disparate fuzzing tools that target different parts of the test environment, such as the scenario, sensors, and vehicle dynamics, there is a lack of fuzzing strategies that ensemble these fuzzers to enable concurrent fuzzing, utilizing diverse techniques and targets. This research proposes FuzzSense, a modular, black-box, mutation-based fuzzing framework that is architected to ensemble diverse AD fuzzing tools. To validate the utility of FuzzSense, a LiDAR sensor fuzzer was developed as a plug-in, and the fuzzer was implemented in the new AD simulation platform AWSIM and Autoware.Universe AD software platform. The results demonstrated that FuzzSense was able to find vulnerabilities in the new Autoware.Universe software. We contribute to FuzzSense open-source with the aim of initiating a conversation in the community on the design of AD-specific fuzzers and the establishment of a community fuzzing framework to better target the diverse technology base of autonomous vehicles.

Paper Structure

This paper contains 22 sections, 6 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. FuzzSense
  3. Fuzzing Broker
  4. Fuzzing Environment
  5. Orchestrator:
  6. Mutator:
  7. Scenario Fuzzer:
  8. Sensor Data Fuzzers:
  9. Oracle and Evaluation:
  10. Repository
  11. Sensor Data Fuzzing
  12. Fuzzing Mask
  13. Multi-Stage Approach
  14. Fuzzing Campaign
  15. Fuzzing Iteration
  16. ...and 7 more sections

Figures (6)

  • Figure 1: High-level Architecture all Components
  • Figure 2: FuzzSense: High-level Architecture of Fuzzing Framework
  • Figure 3: Fuzzing Mask for LiDAR.
  • Figure 4: Fuzzing Mask applied to the right edge of lane
  • Figure 5: Fuzzing Mask applied to central location of vehicle trajectory
  • ...and 1 more figures