Industrial Vehicle Obstacle Classification for Precise Collision Avoidance

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing safety systems for self-propelled industrial vehicles fail to accurately distinguish obstacles, provide refined data for efficient movement control, and suffer from high false alarm rates, leading to reduced operating efficiency.

Innovation Solution

A safety device equipped with depth and RGB video cameras that detect distance and image data, respectively, a control unit that classifies obstacles into categories, and generates differentiated control signals based on obstacle type and distance, allowing precise collision avoidance and alert mechanisms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If video cameras are used to detect images of the surrounding environment, then the system can determine the distance of objects from the vehicle, but it fails to distinguish the type of objects encountered in the trajectory

Engineering Contradiction:
Improvedistance determinationVSAvoidobject type identification
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent segments the detection task into two specialized components: a first video camera dedicated to distance measurement and a second video camera dedicated to object type classification. This segmentation allows each camera to be optimized for its specific function, resolving the contradiction between accurate distance determination and object type identification by distributing these capabilities across separate detection units.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent creates a universal detection system where two video cameras work together to provide both distance measurement and object type classification capabilities. The system integrates multiple functions (distance sensing and object recognition) into a unified safety device, allowing the vehicle to simultaneously obtain both metric information and categorical information about obstacles.

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

2Measurement precision

If depth video cameras are used to accurately determine distances, then the system can locate objects in space, but it fails to determine the type of obstacle

Engineering Contradiction:
Improvedepth measurementVSAvoidobstacle classification
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent divides the detection system into specialized segments: depth video cameras for precise distance measurement and a separate video camera for visual classification of obstacle types. This segmentation enables each component to excel at its specific task without compromise, as the depth camera focuses solely on metric data while the visual camera handles categorical identification.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The control unit acts as an intermediary that integrates data from the depth video cameras and the visual video camera. It combines the precise distance information with the object type classification, creating a comprehensive obstacle understanding that neither camera could achieve alone.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If prior art solutions are used for obstacle detection, then the system can detect objects, but it suffers from high false alarm rates and reduced operating efficiency

Engineering Contradiction:
Improveobstacle detectionVSAvoidoperating efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies different detection qualities to different aspects of obstacle detection: depth video cameras provide high-precision metric quality for distance measurement, while the visual video camera provides categorical quality for object type identification. This local differentiation of detection quality reduces false alarms by providing more nuanced information about each detected object, allowing the system to distinguish between relevant and irrelevant detections.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The control unit processes and integrates information from multiple video cameras, creating a feedback loop that continuously refines obstacle detection accuracy. By combining depth data with visual classification feedback, the system can confirm detections and reduce false alarms, thereby improving reliability without sacrificing operating efficiency.

Inventive Principle:
Principle #23Feedback

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

Enhances obstacle recognition and collision prevention by accurately identifying obstacle types and distances, reducing false alarms and improving operational efficiency through targeted movement control.

Implementation Method 1

determining the distance by determining the return time of the signal sent

Methodology Applied
Scientific EffectTime of flight: Time of Flight

Data Source

PatentEP4204352B1Safety device for self-propelled industrial vehicles
Publication Date: 2026.02.11 KIWITRON SPA
  • EP4204352B1 patent drawingFigure 1
  • EP4204352B1 patent drawingFigure 2
  • EP4204352B1 patent drawingFigure 3

AI summary

Described is a safety device (1) for a self-propelled industrial vehicle (100) movable in an operating space, comprising: a measuring group (11), configured to detect depth data (131), representative of a distance, from the measuring group (11), of bodies arranged in the operating space; an additional measuring group (12), configured to detect image data (132), representative of an image of bodies arranged in the operating space; a control unit (13), connected to the measuring unit (11) and to the additional measuring unit (12) to receive the depth data (131) and the image data (132). The control unit (13) is programmed to compare the image data (132) with reference image data, for classifying the bodies into predetermined categories (CP). The control unit (13) is programmed for associating to each body a corresponding predetermined category (CP) and a corresponding value of distance (Dr) from the self-propelled vehicle.