Stack separation control method, model training method, and stack separation control system
By dynamically adjusting the speed of the belt conveyor unit using a reinforcement learning model, the problem of poor stacking separation effect in existing technologies is solved, achieving efficient and low-cost stacking separation control, which is suitable for express sorting systems.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- WUHAN JIDONG INTELLIGENT TECH CO LTD
- Filing Date
- 2026-04-08
- Publication Date
- 2026-07-17
AI Technical Summary
When faced with packages of varying sizes, weights, and materials, as well as dynamic material feed rates, existing automated conveyor lines cannot make optimal decisions using fixed-rule speed control strategies. This results in poor separation of stacked items and poses high maintenance costs and the risk of equipment damage.
A reinforcement learning model is used to dynamically adjust the speed of the belt conveyor unit. By combining low-dimensional physical features and high-dimensional visual feature data, differentiated speed combinations are achieved through closed-loop control, which reduces system debugging and maintenance costs and improves the separation effect of stacked parts.
It enables adaptive separation of packages of different sizes, weights and shapes, reduces system debugging and maintenance costs, improves separation efficiency, reduces equipment failure risk, and is suitable for the transformation of existing express sorting lines.
Smart Images
Figure CN122403072A_ABST