On a Formal Model of Safe and Scalable Self-driving Cars
arXiv:1708.06374
Abstract
In recent years, car makers and tech companies have been racing towards self driving cars. It seems that the main parameter in this race is who will have the first car on the road. The goal of this paper is to add to the equation two additional crucial parameters. The first is standardization of safety assurance --- what are the minimal requirements that every self-driving car must satisfy, and how can we verify these requirements. The second parameter is scalability --- engineering solutions that lead to unleashed costs will not scale to millions of cars, which will push interest in this field into a niche academic corner, and drive the entire field into a "winter of autonomous driving". In the first part of the paper we propose a white-box, interpretable, mathematical model for safety assurance, which we call Responsibility-Sensitive Safety (RSS). In the second part we describe a design of a system that adheres to our safety assurance requirements and is scalable to millions of cars.
References in corpus (2)
Cited by in corpus (89)
- Criticality Metrics for Automated Driving: A Review and Suitability Analysis of the State of the Art
- Scalable agent alignment via reward modeling: a research direction
- Tackling Occlusions & Limited Sensor Range with Set-based Safety Verification
- F1/10: An Open-Source Autonomous Cyber-Physical Platform
- Cooperative Automated Vehicles: a Review of Opportunities and Challenges in Socially Intelligent Vehicles Beyond Networking
- A Review of Testing Object-Based Environment Perception for Safe Automated Driving
- Software Engineering for AI-Based Systems: A Survey
- Testing and verification of neural-network-based safety-critical control software: A systematic literature review
- Efficient Black-box Assessment of Autonomous Vehicle Safety
- Driving with Style: Inverse Reinforcement Learning in General-Purpose Planning for Automated Driving
- Hierarchical Program-Triggered Reinforcement Learning Agents For Automated Driving
- A Comprehensive Safety Metric to Evaluate Perception in Autonomous Systems
- Deep Reinforcement-Learning-based Driving Policy for Autonomous Road Vehicles
- Formalizing Traffic Rules for Machine Interpretability
- Rigorous Agent Evaluation: An Adversarial Approach to Uncover Catastrophic Failures
- Unsupervised and Supervised Learning with the Random Forest Algorithm for Traffic Scenario Clustering and Classification
- A Survey on Scenario-Based Testing for Automated Driving Systems in High-Fidelity Simulation
- Scalable End-to-End Autonomous Vehicle Testing via Rare-event Simulation
- Worst Cases Policy Gradients
- Finding Critical Scenarios for Automated Driving Systems: A Systematic Literature Review
- Adaptive Stress Testing with Reward Augmentation for Autonomous Vehicle Validation
- A Comparison of Uncertainty Estimation Approaches in Deep Learning Components for Autonomous Vehicle Applications
- Deep Reinforcement Learning for Human-Like Driving Policies in Collision Avoidance Tasks of Self-Driving Cars
- Computation-Communication Trade-offs and Sensor Selection in Real-time Estimation for Processing Networks
- System Design in the Era of IoT --- Meeting the Autonomy Challenge
- Probabilistic Prediction of Vehicle Semantic Intention and Motion
- An Efficient Safety-oriented Car-following Model for Connected Automated Vehicles Considering Discrete Signals
- Towards Safety-Aware Computing System Design in Autonomous Vehicles
- Goal-Aware RSS for Complex Scenarios via Program Logic
- On the validation of complex systems operating in open contexts
- Interaction-aware Decision Making with Adaptive Strategies under Merging Scenarios
- Probabilistic Metamodels for an Efficient Characterization of Complex Driving Scenarios
- Neural Bridge Sampling for Evaluating Safety-Critical Autonomous Systems
- Elements of Effective Deep Reinforcement Learning towards Tactical Driving Decision Making
- Comprehensive Reactive Safety: No Need For A Trajectory If You Have A Strategy
- A Finite-Sampling, Operational Domain Specific, and Provably Unbiased Connected and Automated Vehicle Safety Metric
- Automotive Collision Risk Estimation Under Cooperative Sensing
- Predicting Vehicle Behaviors Over An Extended Horizon Using Behavior Interaction Network
- Vision Zero: on a Provable Method for Eliminating Roadway Accidents without Compromising Traffic Throughput
- Risk-averse Behavior Planning for Autonomous Driving under Uncertainty
- EPSILON: An Efficient Planning System for Automated Vehicles in Highly Interactive Environments
- Risk Structures: Towards Engineering Risk-aware Autonomous Systems
- Comfortable Priority Handling with Predictive Velocity Optimization for Intersection Crossings
- Formal Verification of Safety Architectures for Automated Driving
- Certified Control: An Architecture for Verifiable Safety of Autonomous Vehicles
- A Right-of-Way Based Strategy to Implement Safe and Efficient Driving at Non-Signalized Intersections for Automated Vehicles
- Closing the gap towards end-to-end autonomous vehicle system
- Efficient Uncertainty-aware Decision-making for Automated Driving Using Guided Branching
- Autonomous Vehicles Meet the Physical World: RSS, Variability, Uncertainty, and Proving Safety (Expanded Version)
- Adaptive Stress Testing for Autonomous Vehicles
- Sensors, Safety Models and A System-Level Approach to Safe and Scalable Automated Vehicles
- Corner Case Generation and Analysis for Safety Assessment of Autonomous Vehicles
- Rule-based Optimal Control for Autonomous Driving
- Perceive, Attend, and Drive: Learning Spatial Attention for Safe Self-Driving
- Fault-Tolerant Perception for Automated Driving A Lightweight Monitoring Approach
- A Formally Verified Fail-Operational Safety Concept for Automated Driving
- Self-Driving Vehicle Verification Towards a Benchmark
- AutoPilot: Automating SoC Design Space Exploration for SWaP Constrained Autonomous UAVs
- Teaching AI, Ethics, Law and Policy
- Reliving the Dataset: Combining the Visualization of Road Users' Interactions with Scenario Reconstruction in Virtual Reality
- Active Altruism Learning and Information Sufficiency for Autonomous Driving
- From Sensor to Processing Networks: Optimal Estimation with Computation and Communication Latency
- Universally Safe Swerve Manoeuvres for Autonomous Driving
- Limits of Probabilistic Safety Guarantees when Considering Human Uncertainty
- Inferring Distributions Over Depth from a Single Image
- Generating Socially Acceptable Perturbations for Efficient Evaluation of Autonomous Vehicles
- Learning a Decision Module by Imitating Driver's Control Behaviors
- Autonomics: In Search of a Foundation for Next Generation Autonomous Systems
- Rule-based Evaluation and Optimal Control for Autonomous Driving
- AIBA: An AI Model for Behavior Arbitration in Autonomous Driving
- Driving-Policy Adaptive Safeguard for Autonomous Vehicles Using Reinforcement Learning
- Safety Case Templates for Autonomous Systems
- Representation of Reinforcement Learning Policies in Reproducing Kernel Hilbert Spaces
- Tackling Variabilities in Autonomous Driving
- Efficient falsification approach for autonomous vehicle validation using a parameter optimisation technique based on reinforcement learning
- Tackling Existence Probabilities of Objects with Motion Planning for Automated Urban Driving
- Analysis of draft EU ADS Performance Requirements
- Infusing Reachability-Based Safety into Planning and Control for Multi-agent Interactions
- A taxonomy of strategic human interactions in traffic conflicts
- Verifying Compositional Refinement of Assume/Guarantee Contracts using Linear Programming
- Improving Urban Traffic Throughput with Vehicle Platooning: Theory and Experiments
- Safer End-to-End Autonomous Driving via Conditional Imitation Learning and Command Augmentation
- Oversight of Unsafe Systems via Dynamic Safety Envelopes
- Towards formalization and monitoring of microscopic traffic parameters using temporal logic
- Watch out for the risky actors: Assessing risk in dynamic environments for safe driving
- Discovering Avoidable Planner Failures of Autonomous Vehicles using Counterfactual Analysis in Behaviorally Diverse Simulation
- Quantitative Risk Indices for Autonomous Vehicle Training Systems
- Real-time safety assessment of trajectories for autonomous driving
- Planning on a (Risk) Budget: Safe Non-Conservative Planning in Probabilistic Dynamic Environments