Printer Motor Control via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing printers face challenges in maintaining high positioning accuracy of print media and carriage due to environmental changes and motor deterioration, as current control parameter determination methods are not precise enough.
Innovation Solution
A printer equipped with a machine-learned model that optimizes motor control parameters based on state variables such as speed, acceleration, ambient environment, and print medium type, using reinforcement learning to adjust parameters and improve positioning accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If control parameters are determined by trial and error or fixed reference values, then the device complexity is low, but the positioning accuracy deteriorates over time due to motor deterioration and environmental changes
Solution Approach 1:
The printer performs self-learning by automatically measuring its own transport position using an integrated camera and encoder, and self-adjusts control parameters without external intervention. The system observes state variables, calculates transport position deviations, and updates control parameters autonomously, eliminating the need for manual trial-and-error adjustment while maintaining high positioning accuracy over time.
Solution Approach 2:
The system implements continuous feedback by measuring the actual transport position of the print medium using a camera and encoder, comparing it with the target position, and using this deviation information to adjust control parameters. This closed-loop feedback mechanism enables the system to compensate for motor deterioration and environmental changes, maintaining positioning accuracy without increasing operational complexity.
2Manufacturing precision
If correction techniques are applied to compensate for motor deterioration, then the positioning accuracy is improved, but the ease of operation deteriorates due to required manual adjustments and external measurement devices
Solution Approach 1:
The printer autonomously performs all correction operations without user intervention. The integrated camera captures print medium images, the encoder measures transport position, and the control unit automatically calculates deviations and updates control parameters. This self-service capability eliminates the need for users to perform manual adjustments or connect external measurement devices, maintaining ease of operation while improving positioning accuracy.
Solution Approach 2:
The printer integrates multiple functions into a single system: the camera serves both as a learning tool and a measurement device, the encoder provides both learning data and positioning feedback, and the control unit performs both learning and correction operations. This multi-functionality eliminates the need for separate external measurement devices and manual adjustment procedures, maintaining operational simplicity while achieving high positioning accuracy.
3Adaptability or versatility
If fixed control parameters are used, then the ease of manufacture is high, but the adaptability deteriorates when environmental conditions or motor characteristics change
Solution Approach 1:
The system transitions from static fixed control parameters to dynamic adaptive parameters that automatically adjust to environmental changes and motor deterioration. The printer continuously learns from operational data, updates control parameters based on observed deviations, and adapts to changing conditions. This dynamic approach maintains manufacturing simplicity by implementing learning through software updates rather than hardware changes, while significantly improving adaptability to environmental and operational variations.
Data Source
AI summary
A printer including a motor for transporting an object to be transported, the printer includes: a memory storing a machine-learned model configured to output a control parameter of the motor that brings a transport position of the object to be transported close to a reference based on one or more state variables including at least one of a speed of the object to be transported, an acceleration of the object to be transported, a movement amount of the object to be transported, a start position of movement of the object to be transported, an ambient environment of the printer, a value of a current flowing through the motor, a type of a print medium onto which printing is to be performed by the printer, and an accumulated movement amount of the object to be transported; and a controller configured to control the motor to perform printing by using the control parameter obtained based on the machine-learned model.