Machine-Learning Installation Operations for Lower Call-Out Rates

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing installation processes for people conveyor and access control systems often detect significant issues during the first year post-installation, leading to high call-out rates and operational inefficiencies.

Innovation Solution

A method and system utilizing a computing unit, installation support system, and computer program that leverage site information and machine learning to generate installation operations, including software updates, parameter adjustments, and feature activations, to identify and correct issues in real-time, with the option for remote monitoring and data sharing across installations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional installation processes are used without machine learning, then the installation process is simpler and requires less computational resources, but issue detection capability is insufficient leading to high first-year call-out rates

Engineering Contradiction:
Improveissue detection capabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by training the machine learning model with historical installation data before actual installation processes. The model learns from past issues and installation patterns, enabling it to proactively detect potential problems during ongoing installations before they manifest as actual failures requiring call-outs.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback mechanisms by continuously collecting data from ongoing installation processes and using this feedback to improve the machine learning model's issue detection accuracy. The model receives feedback from detected issues and successful installations, refining its ability to predict and prevent future problems.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If machine learning is applied to detect issues in real-time, then issue detection accuracy improves, but computational resources and processing time increase

Engineering Contradiction:
Improveissue detection accuracyVSAvoidcomputational resource consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The machine learning model is trained in advance using historical installation data from multiple sources including previous installation processes, issue reports, and operational data. This preliminary training enables the model to make accurate predictions during real-time installation monitoring without requiring intensive computational resources during the actual installation process.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system applies partial monitoring and analysis during installation processes, focusing computational resources on detecting specific types of issues that are most critical or commonly occur. Rather than analyzing every possible parameter at maximum depth, the system prioritizes detection of high-risk installation errors.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If site information from multiple installation sites is collected and used for training, then the machine learning model becomes more accurate, but data privacy and security requirements increase

Engineering Contradiction:
Improvemodel training accuracyVSAvoiddata security risks
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The system extracts and utilizes only the necessary installation data and issue patterns from multiple sites while leaving out sensitive proprietary information, customer-specific configurations, and confidential business data. The machine learning model learns general installation patterns and common issues without requiring access to sensitive site-specific information.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces an intermediary layer that processes and anonymizes data from multiple installation sites before feeding it into the machine learning model. This intermediary ensures that raw site information is transformed into generalized patterns that maintain model training accuracy while protecting data privacy and security.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4143679B1A solution for generating at least one installation operation for at least one ongoing installation process at an installation site
Publication Date: 2025.10.22 KONE OYJ
  • EP4143679B1 patent drawingFigure 1~2
  • EP4143679B1 patent drawingFigure 3~4

AI summary

The invention relates to a method for generating at least one installation operation for at least one ongoing installation process of at an installation site (104). The method comprises: obtaining (202) site in- formation of the at least one ongoing installation pro- cess at the installation site (104) from a site control unit (102) arranged at the installation site (104), obtaining (204) site information gathered from one or more previous installation processes at one or more other installation sites from one or more external data- bases (106a-106n), defining (206) at least one installation operation for the at least one ongoing installation process based on the site information of the at least one ongoing installation process and the site information gathered from the one or more previous installation processes, and generating (208) at least one signal comprising an instruction to perform the at least one installation operation for the at least one ongoing installation process to the site control unit (102). The invention relates also to a computing unit (108), an installation support system (100), and a computer program performing at least partly the method.