mlsanityData to 7 Oct 2026

Risk-Driven Design of Perception Systems

arXiv 2205.10677 ↗First posted 21 May 2022cs.RO cs.AIcited by 4

Not on the map. Why?

Abstract (arXiv)

Modern autonomous systems rely on perception modules to process complex sensor measurements into state estimates. These estimates are then passed to a controller, which uses them to make safety-critical decisions. It is therefore important that we design perception systems to minimize errors that reduce the overall safety of the system. We develop a risk-driven approach to designing perception systems that accounts for the effect of perceptual errors on the performance of the fully-integrated, closed-loop system. We formulate a risk function to quantify the effect of a given perceptual error on overall safety, and show how we can use it to design safer perception systems by including a risk-dependent term in the loss function and generating training data in risk-sensitive regions. We evaluate our techniques on a realistic vision-based aircraft detect and avoid application and show that risk-driven design reduces collision risk by 37% over a baseline system.

Key facts. Cited by 4 papers in the map (data through 7 Oct 2026). Builds on 22 of its 50 references that are in the map. First posted on arXiv 21 May 2022.

Builds on

22 of 50 references.

References · 22

  1. 2015You Only Look Once: Unified, Real-Time Object Detectioncited by 4,829
  2. 2017A Distributional Perspective on Reinforcement Learningcited by 743
  3. 2017Distributional Reinforcement Learning with Quantile Regressioncited by 464
  4. 2017Safe Reinforcement Learning via Shieldingcited by 427
  5. 2018Implicit Quantile Networks for Distributional Reinforcement Learningcited by 362
  6. 2015Risk-Sensitive and Robust Decision-Making: a CVaR Optimization Approachcited by 162
  7. 2020Lidar for Autonomous Driving: The principles, challenges, and trends for automotive lidar and perception systemscited by 100
  8. 2014Algorithms for CVaR Optimization in MDPscited by 98
  9. 2020Objective Mismatch in Model-based Reinforcement Learningcited by 87
  10. 2020A Survey of Algorithms for Black-Box Safety Validation of Cyber-Physical Systemscited by 71
  11. 2019Being Optimistic to Be Conservative: Quickly Learning a CVaR Policycited by 63
  12. 2020Adversarial Machine Learning in Image Classification: A Survey Towards the Defender's Perspectivecited by 58
10 more
  1. 2017How Should a Robot Assess Risk? Towards an Axiomatic Theory of Risk in Roboticscited by 57
  2. 2020Backpropagation through Signal Temporal Logic Specifications: Infusing Logical Structure into Gradient-Based Methodscited by 51
  3. 2020Learning to Evaluate Perception Models Using Planner-Centric Metricscited by 49
  4. 2019Adaptive Stress Testing with Reward Augmentation for Autonomous Vehicle Validationcited by 33
  5. 2017Compositional Falsification of Cyber-Physical Systems with Machine Learning Componentscited by 31
  6. 2021Verification of Image-based Neural Network Controllers Using Generative Modelscited by 27
  7. 2017Goal-Driven Dynamics Learning via Bayesian Optimizationcited by 19
  8. 2020Validation of Image-Based Neural Network Controllers through Adaptive Stress Testingcited by 19
  9. 2022Control-Aware Prediction Objectives for Autonomous Drivingcited by 18
  10. 2020Elaborating on Learned Demonstrations with Temporal Logic Specificationscited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 3; 2024: 0; 2025: 0; 2026: 1.

4 papers.

  1. 2023AVOIDDS: Aircraft Vision-based Intruder Detection Dataset and Simulatorcited by 4
  2. 2023Efficient Determination of Safety Requirements for Perception Systemscited by 1
  3. 2023Attacking Motion Planners Using Adversarial Perception Errorscited by 1
  4. 2026Marginal Calibration Does Not Compose: Hidden Dependence in Modular Robot Navigationcited by 0