Autonomous Vehicle Safety Module for Neural Network Control Validation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current autonomous vehicle control systems lack comprehensive safety measures for generating validated control commands, particularly in complex traffic environments, which can lead to unsafe decisions and actions.

Innovation Solution

An end-to-end trained neural network system processes raw sensor data, object-level data, and tactical information to generate control commands, with a safety module performing risk assessment over pre-set time horizons to validate these commands as safe, ensuring holistic decision-making and improved safety.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If an end-to-end trained neural network system is used to generate control commands, then the autonomous vehicle can process information and make decisions in real time, but the system lacks comprehensive safety validation mechanisms leading to unsafe decisions in complex traffic environments

Engineering Contradiction:
Improvereal-time decision-making capabilityVSAvoidsafety validation of control commands
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent introduces a safety module as an intermediary component between the neural network system and the vehicle control system. This safety module validates control commands before execution, acting as a mediator that ensures safety requirements are met while preserving the real-time decision-making capability of the neural network. The safety module checks whether generated control commands satisfy safety constraints without requiring redesign of the entire control architecture.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent implements preliminary safety validation by assessing control commands before they are executed. The safety module performs risk assessment and validation checks in advance, identifying potentially unsafe commands before they reach the vehicle control system. This preliminary action prevents unsafe decisions from being executed while maintaining the efficiency of real-time processing.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If a safety module is added to validate control commands, then safety is improved, but the system complexity increases

Engineering Contradiction:
Improvesafety validation of control commandsVSAvoidcontrol system architecture
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the control system into distinct functional modules: the neural network system for decision-making, the safety module for validation, and the vehicle control system for execution. This segmentation allows each module to perform its specific function independently, making the overall system more manageable and maintainable despite increased complexity. The safety module is designed as a separate, modular component that can be added without fundamentally redesigning the existing neural network architecture.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3588226B1Method and arrangement for generating control commands for an autonomous road vehicle
Publication Date: 2020.06.24 ZENUITY AB
  • EP3588226B1 patent drawingFigure 1
  • EP3588226B1 patent drawingFigure 2
  • EP3588226B1 patent drawingFigure 3

AI summary

Described herein is a method and arrangement (1) for generating validated control commands (2) for an autonomous road vehicle (3). An end-to-end trained neural network system (4) is arranged to receive an input of raw sensor data (5) from on-board sensors (6) of the autonomous road vehicle (3) as well as object-level data (7) and tactical information data (8). The end-to-end trained neural network system (4) is further arranged to map input data (5, 7, 8) to control commands (10) for the autonomous road vehicle (3) over pre-set time horizons. A safety module (9) is arranged to receive the control commands (10) for the autonomous road vehicle (3) over the pre-set time horizons and perform risk assessment of planned trajectories resulting from the control commands (10) for the autonomous road vehicle (3) over the pre-set time horizons. The safety module (9) is further arranged to validate as safe and output validated control commands (2) for the autonomous road vehicle (3).