AGV Motion Control With RLF Zone Prediction Feedback Loops
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AGV systems face reliability issues due to Radio Link Failures (RLF), leading to potential damage or unsafe operation when connection to the control station is lost, with existing technologies lacking effective strategies to predict and manage RLF zones.
Innovation Solution
Implementing a computer-implemented method with a Machine Learning Engine that processes RLF and Position Data through a Mobile Communication System, utilizing RLF and Spatial Information Databases to create a feedback loop for predicting RLF zones, incorporating RSRP/RSRQ and spatial data from on-board devices like LIDAR cameras to enhance prediction accuracy and adaptability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If AGV stops immediately when RLF occurs to ensure safety, then safety is improved, but operational reliability deteriorates because the AGV cannot recover connection and continue operation
Solution Approach 1:
The system performs preliminary actions by predicting RLF zones before the AGV enters them using machine learning models trained on historical RLF data. When an RLF is predicted, the AGV is proactively redirected to alternative routes or safe zones before the connection is lost, rather than waiting for the RLF to occur and then reacting. This allows the AGV to maintain operational continuity while preventing safety issues.
Solution Approach 2:
The system implements a feedback mechanism where RLF occurrence data and position data are continuously collected, stored in databases, and used to retrain and improve the machine learning prediction models. This closed-loop feedback enables the system to learn from actual RLF events and improve its prediction accuracy over time, allowing progressively more reliable operation in previously problematic zones.
2Ease of operation
If AGV operates in one-dimensional pre-defined routes to simplify control, then ease of operation is improved, but reliability deteriorates due to limited access to mobile communication services
Solution Approach 1:
The system transitions from one-dimensional pre-defined linear routes to two-dimensional operational domains by enabling the AGV to move laterally between multiple possible paths. The machine learning model predicts RLF zones across the two-dimensional space and directs the AGV to navigate around these zones using alternative routes, thereby maintaining connection stability while preserving operational simplicity through automated guidance.
3Reliability
If machine learning engine processes RLF and position data to predict RLF zones, then reliability is improved, but device complexity increases
Solution Approach 1:
The system introduces intermediary components including databases for storing RLF and position data, and machine learning engines for processing this data into predictions. These intermediaries act as dedicated specialists that handle the complex data processing tasks, allowing the AGV control system to remain relatively simple while benefiting from sophisticated prediction capabilities provided by the intermediary ML components.
Data Source
AI summary
A computer-implemented method (CIM) is disclosed for controlling the motion of at least one AGV (AGV), comprising sending RLF Zone Prediction Data to this at least one AGV (AGV) via a Mobile Communication System (MCS), sending RLF and Position Data via the Mobile Communication System (MCS) to a RLF Information Database (RDB) which is connected to a Machine Learning Engine (MLE), the Machine Learning Engine (MLE) processing the RLF and Position Data to RLF Zone Prediction Data that is sent back to the Mobile Communication System (MCS) and completes a RLF Feedback Loop (I). In order to minimize the impact of RLFs on a system of AGVs the computer-implemented method (CIM) comprises the additional steps of - receiving Position-dependent Spatial and RSRP/RSRQ Data from the at least one AGV (AGV) via the Mobile Communication System (MCS) and sending that data to a Spatial Data Merger (SDM) that connects the Position-dependent Spatial and RSRP/RSRQ Data, and adding this Position-dependent Spatial and RSRP/RSRQ Data to a Spatial Information Database (SDB), - employing the Machine Learning Engine (MLE) that computes the RLF Zone Prediction Data out of entries of the RLF Information Database (RDB) and the Spatial Information Database (SDB) - sending back this RLF Zone Prediction Data to the Mobile Communication System (MCS) and - completing a Spatial Feedback Loop (II).