Autonomous Vehicle Observability Control for Reliable Object Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Autonomous vehicles face challenges in ensuring the reliability and safety of autonomous driving due to limitations in sensor data quality and environmental observability, particularly in adverse weather conditions or sensor obstructions, which can lead to missed object detection and false alarms.

Innovation Solution

A control unit determines an observability value for each location in the vehicle's environment, indicating how well that area can be sensed by multiple sensors, and adjusts vehicle operations such as speed and trajectory based on this value, using an observability grid to account for sensor capabilities and weather conditions, thereby enhancing safety and reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If sensor data is used to detect objects in the environment, then object detection capability is improved, but reliability decreases under adverse weather conditions or sensor obstructions

Engineering Contradiction:
Improveobject detection capabilityVSAvoiddetection reliability
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The system pre-determines an observability value for each location in the environment before actual object detection occurs. This observability value, stored in an observability grid, represents the expected sensor coverage and detection quality for that location under current environmental conditions. By having this information prepared in advance, the system can proactively adjust its operation plan to avoid locations with poor observability, thereby preventing unreliable detections rather than reacting to them after they occur.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system continuously determines observability values based on current sensor data and environmental conditions, updating the observability grid in real-time. This feedback loop allows the system to monitor the quality of sensor coverage across the environment and adjust its operation plan dynamically. The control unit uses this feedback to modify the vehicle's trajectory or speed to maintain operation only in areas where object detection remains reliable, thus resolving the contradiction between detection capability and reliability under varying conditions.

Inventive Principle:
Principle #23Feedback

2Productivity

If the vehicle operates at higher speed to increase productivity, then output is improved, but safety decreases due to reduced sensor performance

Engineering Contradiction:
Improvevehicle outputVSAvoidsafety
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system dynamically adjusts the vehicle's operating parameters including speed and trajectory based on real-time observability conditions. The control unit continuously monitors the observability grid and modifies the operation plan to maintain the vehicle only in areas where sensor coverage and object detection reliability meet safety thresholds. This dynamic adaptation allows the system to optimize productivity by operating at higher speeds in areas with good observability while automatically reducing speed or changing trajectory when entering areas with poor sensor performance, thus maintaining safety without completely sacrificing productivity.

Inventive Principle:
Principle #15Dynamics

3Adaptability or versatility

If the vehicle operates in areas with poor observability, then adaptability is improved, but measurement precision decreases

Engineering Contradiction:
Improveoperational flexibilityVSAvoidobject detection accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system pre-assesses the observability of different locations in the environment using sensor data and environmental information, creating an observability grid that identifies areas with sufficient versus insufficient sensor coverage. The control unit uses this pre-determined information to proactively plan the vehicle's operation, selecting trajectories and locations where observability meets minimum thresholds for reliable object detection. This prevents the vehicle from entering areas where measurement precision would be compromised, while still providing operational flexibility by dynamically adjusting the plan based on current conditions.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3901661B1Method and control unit for operating an autonomous vehicle
Publication Date: 2024.05.22 BAYERISCHE MOTOREN WERKE AG
  • EP3901661B1 patent drawingFigure 1~2a
  • EP3901661B1 patent drawingFigure 2b
  • EP3901661B1 patent drawingFigure 2c

AI summary

The present document describes a control unit (101) for operating a vehicle (100) which comprises a plurality of sensors (102) for capturing sensor data regarding an environment of the vehicle (100). The control unit (101) is configured to determine an observability value (301) for a particular location (201) within the environment of the vehicle (100); wherein the observability value (301) is indicative of how well the particular location (201) can be sensed by the plurality of sensors (102) of the vehicle (100). Furthermore, the control unit (101) is configured to operate the vehicle (100) in dependence of the observability value (301) for the particular location (201).