Robot THT PCB Insertion Using RL Visual Feedback

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional methods for inserting electronic components into printed circuit boards using industrial robots are inefficient due to complex geometry and susceptibility to damage, requiring high precision sensors and manual parameter tuning, and are limited by compliance control systems that perform preprogrammed trajectories.

Innovation Solution

The method employs reinforcement learning (RL) algorithms, specifically the Soft Actor-Critic (SAC) algorithm integrated with an admittance control system and convolutional neural networks (CNNs) for image processing, using a camera-mounted end-effector to calculate movement instructions and adjust the tool's position based on real-time image feedback and sensor data, allowing for precise and adaptive insertion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If conventional compliance control systems with preprogrammed trajectories are used, then the robot can perform insertion tasks, but the method requires high precision sensors and manual parameter tuning which is time-consuming

Engineering Contradiction:
Improveautomation of insertion taskVSAvoidcomplexity of control system
Core Design Contradiction:
Extent of automationVSDevice complexity

Solution Approach 1:

The patent replaces the conventional compliance control system with reinforcement learning algorithms that process visual feedback from cameras. Instead of using complex force control systems with high-precision sensors, the invention uses deep learning-based visual compliance to guide the robot's insertion movements, substituting mechanical control with intelligent software-based control.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The reinforcement learning system enables the robot to autonomously learn and optimize insertion parameters through trial and error, eliminating the need for manual parameter tuning. The system self-adjusts by processing visual feedback and automatically optimizing its control strategy, making the commissioning process faster and more autonomous.

Inventive Principle:
Principle #25Self-service

2Reliability

If conventional methods with constant downforce are used, then the robot can maintain contact with the component, but the method is limited by preprogrammed trajectories and cannot adapt to complex geometries

Engineering Contradiction:
Improvecontact maintenanceVSAvoidadaptability to complex geometries
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic adaptation by using reinforcement learning to continuously adjust insertion parameters based on real-time visual feedback. Instead of fixed preprogrammed trajectories, the system dynamically optimizes the insertion path and forces applied, allowing adaptation to various component geometries and positions while maintaining reliable contact.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system uses visual feedback from cameras mounted on the robot end-effector to continuously monitor the insertion process. This feedback loop enables the reinforcement learning algorithm to adjust control parameters in real-time, maintaining reliable contact while adapting to complex geometries that cannot be handled by preprogrammed trajectories.

Inventive Principle:
Principle #23Feedback

3Manufacturing precision

If high precision sensors and well-made robotized production machines are used, then accurate insertion can be achieved, but the system requires manual parameters tuning which reduces productivity

Engineering Contradiction:
Improveinsertion accuracyVSAvoidproduction speed
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The reinforcement learning system automatically tunes insertion parameters through autonomous learning, eliminating the time-consuming manual parameter adjustment process. The system self-optimizes by processing visual feedback and learning from trial and error, achieving accurate insertion without sacrificing production speed through automated parameter optimization.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces manual parameter tuning with automated reinforcement learning algorithms that process visual information and automatically optimize insertion parameters. This substitution of manual operations with intelligent automation maintains manufacturing precision while significantly improving productivity by eliminating time-consuming parameter adjustment processes.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

4Ease of manufacture

If preprogrammed trajectories are used, then the control system is simple to implement, but the method cannot adapt to variations in component position and geometry

Engineering Contradiction:
Improveease of control system implementationVSAvoidadaptability to position variations
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The system implements visual feedback using cameras mounted on the robot end-effector to detect component position and geometry variations in real-time. This feedback enables the reinforcement learning algorithm to adapt the insertion trajectory dynamically, maintaining ease of implementation through standard robot hardware while achieving high adaptability to position and geometry variations.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent transforms the static preprogrammed trajectory into a dynamic adaptive path using reinforcement learning. The system maintains simplicity by using standard robot control interfaces while achieving adaptability through intelligent algorithms that process visual feedback and automatically adjust the insertion trajectory to accommodate variations in component position and geometry.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4338895A1A method of inserting an electronic component in through-hole technology, THT, into a printed circuit board, PCB, by an industrial robot
Publication Date: 2024.03.20 FITECH SPOKA Z OGRANICZON ODPOWIEDZIALNOCI
  • EP4338895A1 patent drawingFigure 1
  • EP4338895A1 patent drawingFigure 2
  • EP4338895A1 patent drawingFigure 3

AI summary

The object of the invention is the method of inserting an electronic components in through-hole technology, THT, into a printed circuit board, PCB 3, by an industrial robot 1. This method is based on reinforcement learning (RL) settings. A method of inserting an electronic components in through-hole technology, THT, into a printed circuit board, PCB 3, by an industrial robot 1, comprising steps of: a) grabbing, by means of a tool with universal fingers mounted to the end-effector of the industrial robot 1, the electronic component to be inserted into the PCB 3; b) moving the tool to a starting position being in close proximity to a final position of the electronic component; c) acquiring at least one image showing the tool, the electronic component and the PCB 3; d) calculating, on a basis of the at least one image, at least one movement instruction for the industrial robot 1; e) adjusting position of the tool on a basis of the at least one movement instruction; f) repeating steps c), d), and e) until the electronic component is in the final position.