Remote Control Delay Compensation Using Reinforcement Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing control systems face challenges in maintaining optimal control operations due to unstable communication delays between remote and local control apparatuses, leading to fluctuations in control delay periods that affect the effectiveness of control operations.
Innovation Solution
A control system that includes a remote-control apparatus and a local-control apparatus, utilizing machine learning to generate a model that accounts for communication delays, enabling the calculation of control values that compensate for these fluctuations, thereby ensuring optimal control operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If remote control is implemented over communication networks, then operational flexibility and distance independence are improved, but control delay fluctuations and stability deteriorate
Solution Approach 1:
The system performs preliminary actions by calculating multiple candidate control values in advance, each corresponding to different delay amounts. The remote control apparatus computes control values for various possible delay scenarios before the actual control execution, so when the control value is transmitted and applied, the system is already prepared with pre-calculated options that account for potential delay variations.
Solution Approach 2:
The system makes the control system dynamic by adapting control values based on actual measured delay amounts. Instead of using a fixed control value, the local control apparatus measures the actual communication delay and selects or adjusts the control value to match the observed delay conditions, allowing the system to dynamically respond to varying network conditions.
2Device complexity
If control values are calculated without considering delay fluctuations, then calculation simplicity is improved, but control effectiveness deteriorates
Solution Approach 1:
The control calculation process is segmented into multiple independent calculations, where the remote control apparatus calculates multiple candidate control values corresponding to different delay amounts separately. Each candidate control value is computed independently based on different assumed delay conditions, allowing the system to handle complexity through division rather than requiring a single complex calculation that accounts for all possible delays simultaneously.
3Adaptability or versatility
If multiple control values are calculated for different delay amounts, then control adaptability is improved, but computational load increases
Solution Approach 1:
The system creates copies of control values for different delay scenarios rather than calculating a single optimal control value. The remote control apparatus generates multiple candidate control values that are copies or variations of what would be a single control value, each tailored to specific delay conditions. This allows the system to adapt to different delay amounts without requiring complex real-time calculations during control execution.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
[Solving means] A remote-control apparatus, including a measurement value receiving unit for receiving a measurement value related to equipment from a local-control apparatus configured to control the equipment; a calculating unit for calculating a control value corresponding to the measurement value received by the measurement value receiving unit and a delay amount, by using a model configured to calculate a control value that should be used for control of the equipment when there is caused control delay including communication delay with the local-control apparatus from a delay amount corresponding to the control delay and a measurement value; and a control value transmitting unit for transmitting the control value calculated by the calculating unit to the local-control apparatus, is provided.