Sensor Fusion Grid for Obstacle Detection Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing sensor systems for vehicles, such as radar and sonar, often produce false positives and misdetections due to noise and clutter, leading to inefficient obstacle avoidance and mission interference, especially when the vehicle's absolute position is unknown.

Innovation Solution

A method and system for sensor fusion that combines filtered readings from multiple sensors, like radar and ultrasonic systems, to create a composite grid referenced to a vehicle coordinate frame, which is updated based on vehicle movement or regular intervals, allowing for accurate obstacle detection and vehicle guidance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If radar or sonar systems are individually used to detect obstacles, then the system can operate independently with each sensor, but false positives and misdetections occur due to noise and clutter

Engineering Contradiction:
Improveobstacle detection accuracyVSAvoidfalse positives and noise
Core Design Contradiction:
ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent combines data from multiple sensors (radar, sonar, LIDAR, cameras) to create a unified obstacle detection system. By merging sensor inputs, the system cross-validates detections and reduces false positives that would occur with individual sensors, directly improving reliability while filtering out noise and clutter through multi-sensor correlation.

Inventive Principle:
Principle #5Merging (Combining)

2Reliability

If multiple sensors are combined to improve detection accuracy, then false positives are reduced, but the device complexity increases

Engineering Contradiction:
Improveobstacle detection accuracyVSAvoidsensor fusion system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the sensor fusion process into distinct functional modules: individual sensor data processing, coordinate transformation, data association, and obstacle detection. This modular segmentation manages complexity by organizing the multi-sensor system into manageable, independent processing stages that can be implemented and maintained separately while achieving reliable obstacle detection.

Inventive Principle:
Principle #1Segmentation

3Adaptability or versatility

If the vehicle's absolute position is unknown, then the system can operate without GPS, but the accuracy of obstacle position determination deteriorates

Engineering Contradiction:
Improveoperation without global positioningVSAvoidobstacle position accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent introduces relative positioning as an intermediary approach when absolute GPS positioning is unavailable. The system uses the vehicle's own motion and sensor data to establish relative positions of obstacles with respect to the vehicle, maintaining operational capability without global positioning while achieving sufficient precision for obstacle avoidance through local coordinate transformations and motion compensation.

Inventive Principle:
Principle #24Intermediary (Mediator)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enhances the accuracy of obstacle detection and vehicle navigation by reducing false positives and noise, enabling efficient obstacle avoidance and mission execution even without certain global positioning, and is suitable for both manned and unmanned vehicles.

Implementation Method 1

The first sensor 10 comprises a radar system

Methodology Applied
Scientific EffectRadar: Radar

Implementation Method 2

the second sensor 14 comprises an ultrasonic system

Methodology Applied
Scientific EffectUltrasonic: Ultrasonic Vibration

Data Source

PatentEP1742083B1Method and system for sensor signal fusion
Publication Date: 2013.01.02 DEERE & CO
  • EP1742083B1 patent drawingFigure 1
  • EP1742083B1 patent drawingFigure 2
  • EP1742083B1 patent drawingFigure 3

AI summary

A first sensor (10) collects a first group of first sensor readings of first fixed size dependent on a range of a first sensor (10) and safeguarding requirements of a vehicle. A second sensor (14) collects a second group of second sensor readings of second fixed size dependent on a range of a second sensor and the safeguarding requirements. A reference frame manager references the first group and the second group of readings to a vehicle coordinate reference frame. An integration module establishes a composite grid referenced to the ground based on a spatially aligned integration of the readings of the first group and the second group. An update module refreshes the composite grid upon material movement of the vehicle (702) such that the global occupancy of each cell in the composite grid varies in accordance with the particular location of the vehicle (70) on the terrain. The composite grid can be used for automatically guiding the vehicle.