SCADA Radio Access Scheduling Using Readiness Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing SCADA systems face challenges in achieving high throughput and reducing latency in remote terminal communications due to the inability of access points to predict which remote data radios are ready to transmit data, leading to inefficient channel allocation.
Innovation Solution
Implementing reinforced machine learning algorithms at the access point to analyze message headers and predict response times of remote data radios, enabling a centrally coordinated access protocol that allocates channel time efficiently to radios ready to transmit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a random access protocol is used for channel access, then low latency is achieved under low traffic conditions, but throughput is limited to approximately 30% of channel capacity under high traffic conditions
Solution Approach 1:
The system performs preliminary actions by using machine learning to predict which data radios will be ready to transmit in future time slots before actually allocating channels. This allows the access point to pre-prepare channel allocations based on predicted readiness, avoiding the need to wait for random access requests while ensuring channels are allocated only to radios actually ready to transmit, thus achieving both low latency and high throughput under high traffic conditions
2Productivity
If a centrally coordinated access protocol is used, then near 100% of channel capacity can be achieved, but the access point must have knowledge of when radios are ready with data to transmit which is generally unavailable
Solution Approach 1:
The system implements feedback by using machine learning models that continuously learn from observed patterns in data radio behavior. The access point receives feedback in the form of message transmissions and uses this information to refine predictions about which radios will be ready to transmit in future time slots, enabling centralized coordination without requiring direct knowledge of radio readiness states
Solution Approach 2:
The patent replaces the mechanical/informational requirement of directly knowing radio readiness states with a computational prediction system. Instead of requiring actual knowledge of when radios are ready (informational mechanism), the system uses machine learning algorithms to computationally predict readiness based on observed patterns, substituting direct information with intelligent inference
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
Methods and systems for managing remote terminal communications in remote Supervisory Control and Data Acquisition (SCADA) and telemetry networks. Reinforced machine learning processes header data of multiple messages received from remote data radios to predict which of the data radios are currently ready with a response to transmit based on or using information associated with a learned time delay. Time on shared wireless channels is efficiently allocated to the data radios that are currently ready with the response.