Generating Realistic and Diverse Tests for LiDAR-Based Perception Systems
Autonomous systems rely on a perception component to interpret their surroundings, and when misinterpretations occur, they can and have led to serious and fatal system-level failures. Yet, existing methods for testing perception software remain limited in their capacity to efficiently generate test data that is both realistic in order to translate to real-world performance and diverse in order to capture the long tail of rare but safety-critical scenarios. These limitations are particularly evident for perception systems based on LiDAR sensors, which have emerged as a crucial component in modern autonomous systems due to their ability to provide a 3D scan of the world and operate in all lighting conditions. To address these limitations, we introduce a novel approach for testing LiDAR-based perception systems by leveraging existing real-world data as a basis to generate realistic and diverse test cases through mutations that preserve realism invariants while generating inputs rarely found in existing data sets, and automatically crafting oracles that identify potentially safety-critical issues in perception performance. We implemented our approach to assess its ability to identify perception failures, generating over 50,000 test inputs for five state-of-the-art LiDAR-based perception systems. We found that it efficiently generated test cases that yield errors in perception that could result in real consequences if these systems were deployed and does so at a low rate of false positives.
Fri 19 MayDisplayed time zone: Hobart change
15:45 - 17:15 | Cyber-physical systems testingSEIP - Software Engineering in Practice / Technical Track / Journal-First Papers at Meeting Room 106 Chair(s): Shahar Maoz Tel Aviv University | ||
15:45 15mTalk | Data-driven Mutation Analysis for Cyber-Physical Systems Journal-First Papers Enrico ViganĂ² University of Luxembourg, Oscar Cornejo SnT Centre, University of Luxembourg, Fabrizio Pastore University of Luxembourg, Lionel Briand University of Luxembourg; University of Ottawa Link to publication Pre-print | ||
16:00 15mTalk | Finding Causally Different Tests for an Industrial Control System Technical Track Chris Poskitt Singapore Management University, Yuqi Chen ShanghaiTech University, China, Jun Sun Singapore Management University, Yu Jiang Tsinghua University DOI Pre-print File Attached | ||
16:15 15mTalk | Doppelganger Test Generation for Revealing Bugs in Autonomous Driving Software Technical Track Yuqi Huai University of California, Irvine, Yuntianyi Chen University of California, Irvine, Sumaya Almanee University of California, Irvine, Tuan Ngo VNU University of Engineering and Technology, Xiang Liao University of California, Irvine, Ziwen Wan University of California, Irvine, Qi Alfred Chen University of California, Irvine, Joshua Garcia University of California, Irvine Pre-print | ||
16:30 15mTalk | Generating Realistic and Diverse Tests for LiDAR-Based Perception Systems Technical Track Garrett Christian University of Virginia, Trey Woodlief University of Virginia, Sebastian Elbaum University of Virginia Pre-print | ||
16:45 15mTalk | Automated Test Case Generation for Safety-Critical Software in Scade SEIP - Software Engineering in Practice Elson Kurian University of Milano Bicocca, Pietro Braione University of Milano-Bicocca, Daniela Briola University of Milano Bicocca, Dario D'Avino , Matteo Modonato , Giovanni Denaro University of Milano-Bicocca, Italy | ||
17:00 7mTalk | Single and Multi-objective Test Cases Prioritization for Self-driving Cars in Virtual Environments Journal-First Papers Christian Birchler Zurich University of Applied Sciences, Sajad Khatiri USI-Lugnao & Zurich University of Applied Sciences, Pouria Derakhshanfar JetBrains Research, Sebastiano Panichella Zurich University of Applied Sciences, Annibale Panichella Delft University of Technology | ||
17:07 7mTalk | Parameter Coverage for Testing of Autonomous Driving Systems Under Uncertainty Journal-First Papers Thomas Laurent JSPS@National Institute of Informatics, Japan, Stefan Klikovits Johannes Kepler University, Linz, Paolo Arcaini National Institute of Informatics
, Fuyuki Ishikawa National Institute of Informatics, Anthony Ventresque Trinity College Dublin & Lero, Ireland Link to publication DOI |