Conveyor Merge Control Using Learned Drive Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complex control of multiple parallel conveyor lines in logistics systems, requiring dynamic optimization of throughput while avoiding malfunctions, is challenging due to the need for manual adjustments and varying package sizes and mechanical structures.
Innovation Solution
A computer-implemented method using machine learning, specifically supervised learning and reinforcement learning, to determine configuration data for partial conveyor sections, enabling automatic control signal generation for optimal package alignment and throughput on a single output conveyor line.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple parallel conveyor lines are used to increase throughput, then productivity is improved, but device complexity increases due to the need for independent control of each sub-conveyor line
Solution Approach 1:
The patent combines multiple independent conveyor line controls into a single integrated control system. The computing unit centrally manages all sub-conveyor lines, receiving sensor data from the entire system and coordinating acceleration/deceleration commands across all lines simultaneously. This merging of control functions reduces the complexity that would arise from multiple separate control systems while maintaining the productivity benefits of parallel conveyor lines.
Solution Approach 2:
The control system is designed with universal functionality to handle different mechanical configurations (varying numbers of conveyor lines, different sub-conveyor line lengths) through a single adaptable controller. The system can optimize throughput for various package sizes and properties using the same control architecture, eliminating the need for specialized control systems for each configuration scenario.
2Adaptability or versatility
If the control system is adapted to different mechanical configurations (number of conveyor lines, length of sub-conveyor lines), then adaptability is improved, but device complexity increases due to the impact on the control process
Solution Approach 1:
The control system employs dynamic parameters that can be adjusted based on the actual mechanical configuration. The computing unit receives real-time data about the number of conveyor lines, sub-conveyor line lengths, and package characteristics, then dynamically optimizes acceleration and deceleration profiles. This dynamic adaptation allows the system to handle varying configurations without requiring complex reconfiguration of the control architecture.
Solution Approach 2:
The system changes control parameters (acceleration values, deceleration values, timing) based on detected mechanical configurations and package properties. By adjusting these parameters dynamically rather than redesigning the control system for each configuration, the patent achieves high adaptability while keeping the control process relatively simple and unified.
3Productivity
If dynamic optimization of throughput is implemented during operation, then productivity is improved, but reliability worsens due to the risk of malfunctions and collisions
Solution Approach 1:
The system continuously receives feedback from sensors that detect package positions, conveyor line speeds, and system state. The computing unit uses this real-time feedback to adjust acceleration and deceleration commands, ensuring that throughput optimization does not compromise safety. The feedback loop allows the system to respond to actual conditions rather than following pre-programmed sequences, preventing collisions while maintaining high throughput.
Solution Approach 2:
The control system performs preliminary calculations and preparations for acceleration and deceleration maneuvers before executing them. By planning control actions in advance based on predicted package positions and system state, the system can optimize throughput while ensuring that safety constraints are met before dynamic changes are implemented, rather than reacting to potential conflicts after they arise.
Data Source
Figure 1
Figure 2
AI summary
The invention describes a method for computer-implemented configuration of a controlled drive application of a logistics system (1). The logistics system (1) comprises a plurality of parallel conveying paths (10, 20, 30) for piece goods, which paths each lead in the conveying direction (FR) into a merging unit (40). Each of the conveying paths (10, 20, 30) consists of a plurality of sub-conveying paths (11-13, 21-23, 31-33) which, controlled by a computing unit (60), are each accelerated or delayed by their own dedicated drive (11A-13A, 21A-23A, 31A-33A) in order to enable the merging unit (40) to merge the piece goods on a single output conveying path (50) with defined spacing. A system model of the logistics system (1) is firstly determined on the basis of operating data of the logistics system (1) which are present for multiple times in the operation of the logistics system and which comprise sensor values of the logistics system and changes to control variables. On the basis of the system model, a control function of the logistics system (1) is then determined, which comprises at least configuration data (KD) for the drives (11A-13A, 21A-23A, 31A-33A), with at least one control action being performed on the precondition of one or more performance features that are to be achieved in the system model, during which control action the operating data (BD) is simulated for a plurality of time steps. A reward measure is determined for each time step, with the control action being used as a control function for which a specified fitness function, which aggregates the reward measures of a plurality of time steps, satisfies a predetermined criterion.