Vehicle Sensor Validation Using Cross-Sensor Velocity Consistency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for validating the extrinsic calibration of environmental sensors on vehicles are inadequate, particularly in non-overlapping sensor coverage areas and unmapped environments, and often require object identification and digital maps for validation.

Innovation Solution

A method that determines a homogeneous coordinate transformation to convert sensor coordinates into a vehicle coordinate system, identifies deviations in object velocities detected by multiple sensors, and uses a motion model to detect decalibrated states without requiring overlapping sensor coverage or digital maps.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If existing validation methods are used, then validation can be performed in overlapping sensor coverage areas with digital maps, but validation cannot be performed in non-overlapping coverage areas or unmapped environments

Engineering Contradiction:
Improvevalidation applicabilityVSAvoidvalidation reliability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The sensor system validates itself by using its own measurements from multiple sensors to detect decalibrated states, eliminating the need for external digital maps or overlapping coverage areas. The system serves its own validation needs through internal consistency checks of velocity measurements across different sensors.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The validation method works universally across all sensor coverage scenarios including non-overlapping areas and unmapped environments, not just in limited conditions. The same validation approach using velocity comparisons can be applied regardless of sensor geometry or environmental mapping status.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Measurement precision

If object identification and digital maps are required for validation, then validation accuracy may be improved in specific cases, but the system complexity and requirements increase

Engineering Contradiction:
Improvevalidation precisionVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The method extracts and uses only the velocity measurement information from multiple sensors, removing the need for complex object identification and digital map infrastructure. By focusing on the fundamental velocity data that all sensors provide, the system achieves validation without additional complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

3Productivity

If extrinsic calibration is not validated, then sensor system operation continues uninterrupted, but erroneous vehicle functions may occur due to decalibrated sensors

Engineering Contradiction:
Improvesystem continuityVSAvoidfunction accuracy
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system continuously monitors velocity measurements from multiple sensors and provides feedback about the consistency of these measurements. When decalibrated states are detected through velocity comparisons, the system can trigger recalibration or alert operators, creating a closed-loop feedback mechanism that maintains long-term accuracy.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4260082B1Method for validating surroundings detection sensors of a vehicle, and vehicle designed to validate surroundings detection sensors
Publication Date: 2025.12.17 DAIMLER TRUCK AG
  • EP4260082B1 patent drawingFigure 1
  • EP4260082B1 patent drawingFigure 2
  • EP4260082B1 patent drawingFigure 3

AI summary

The invention relates to a method for validating a plurality of surroundings detection sensors, which are rigidly connected to a vehicle (1), for respectively detecting a speed (V1 to V7, Vn) relative to a sensor coordinate system (S1 to S7, Sn). In an intrinsic calibration process, a homogeneous coordinate transformation is determined for each sensor coordinate system (S1 to S7, Sn) with respect to a vehicle coordinate system (V) rigidly connected to the vehicle (1). On the basis of the coordinate transformation, an object speed (X1 to X7, Xn) relative to the vehicle coordinate system (V) and/or parameters of a movement model of the vehicle (1) are determined at each relative speed (V1 to V7, Vn). An uncalibrated state (C0) is then assigned if the object speeds (X1 to X7, Xn) deviate from one another and/or with respect to the movement model of the vehicle (1) by more than a specified degree. The invention additionally relates to a vehicle (1) comprising surroundings detection sensors and at least one computing unit for carrying out such a method.