{"id":132657,"date":"2025-11-19T10:57:26","date_gmt":"2025-11-19T02:57:26","guid":{"rendered":"https:\/\/www.tm-robot.com\/docs\/intelligent-cloning-workflow\/"},"modified":"2026-01-22T14:09:16","modified_gmt":"2026-01-22T06:09:16","slug":"intelligent-cloning-workflow","status":"publish","type":"docs","link":"https:\/\/www2.tm-robot.com\/en\/docs\/intelligent-cloning-workflow\/","title":{"rendered":"Intelligent Cloning Workflow"},"content":{"rendered":"

TMflow Software version: 2 Series<\/p>\n

TM Robot Hardware version: All Robot with EIH Camera<\/p>\n

Jig: A fixture designed to be mounted on the FOUP with a tolerance of \u00b10.05 mm, and capable of supporting LM attachment.<\/p>\n

\n

Background<\/h1>\n

In Robot + AMR integrated applications, system integrators typically use the Landmarks (LM) attached to each slot on the E-Rack to establish the LM Base. This LM Base is then used as the reference for teaching the FOUP pick-and-place points.<\/p>\n

Ideally, once the teaching is completed, the same LM Base should be applicable to all slots. However, in actual operation, the following factors often introduce variations:<\/p>\n

A. Nonlinear errors may occur when the robot performs Cartesian movements in different postures.<\/p>\n

B. Inconsistencies in LM attachment positions or orientations across slots can also lead to pick-and-place deviations.<\/p>\n

These errors may cause the robot\u2019s actual physical position to deviate when moving to the intended pick-and-place points, potentially resulting in mechanical interference.<\/p>\n

At present, most system integrators still rely on manually fine-tuning the offsets for each slot individually. This process is time-consuming, labor-intensive, and in urgent need of a more efficient solution.<\/p>\n

\"\"<\/a>
Figure 1 : Typical Approach<\/figcaption><\/figure>\n

Introduction & Value<\/h1>\n

This method effectively resolves issue B<\/strong>, namely the pick-and-place errors caused by inconsistencies in the LM attachment positions across slots. With this approach, when the robot performs pick-and-place operations at each slot under similar postures, no additional manual fine-tuning is required. The pick-and-place accuracy can be maintained within \u00b11 mm<\/strong>, significantly reducing the amount of manual work required.<\/p>\n

For example:
\nSuppose there are 100 slots on site. With the traditional approach, each slot requires individual teaching and fine-tuning, taking approximately 10 minutes per slot\u2014resulting in a total of 100 \u00d7 10 = 1,000 minutes<\/strong>.<\/p>\n

With the new method, the workflow becomes:
\n\u2022 Initial pick-and-place teaching: approximately 10 minutes
\n\u2022 For all remaining slots, only 1 minute per slot is needed to run the vision task and establish the required relationship<\/p>\n

In total, only 10 + (1 \u00d7 100) = 110 minutes<\/strong> are needed to complete the pick-and-place point setup for all slots\u2014saving nearly 90%<\/strong> of the time compared to the traditional process while also reducing the likelihood of human error.<\/p>\n

Preparation<\/h1>\n

TMflow Software version: 2 Series<\/p>\n

TM Robot Hardware version: All Robot with EIH Camera<\/p>\n

Jig: A fixture designed to be mounted on the FOUP with a tolerance of \u00b10.05 mm, and capable of supporting LM attachment.<\/p>\n

\"\"
Figure 2 : Jig Design<\/figcaption><\/figure>\n

Optimization Method & Process<\/h1>\n