Predictive Latency Compensation for Real-Time Client Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication technologies for cloud and edge computing in industrial environments, particularly TCP/IP-based connections, fail to support real-time capabilities and deterministic behavior, leading to unpredictable latencies that can negatively impact the performance of real-time critical software functions in mobile devices and machines.
Innovation Solution
A method for latency compensation involving a client device that receives predictions from a server device, detects sensor signals, waits for a predetermined time for operation commands, and selects the most accurate prediction to operate based on sensor signal hypotheses, ensuring real-time capability even with varying network latencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If TCP/IP-based connections are used for cloud/edge computing, then computing power can be outsourced flexibly and scalably, but real-time capability and deterministic behavior are lost due to unpredictable latencies
Solution Approach 1:
The server generates multiple predictions in advance, each with sensor signal hypotheses and operation command hypotheses for future time steps. These predictions are sent to the client before the actual execution time, so when latency occurs, the client already has pre-computed commands to fall back on.
Solution Approach 2:
The patent introduces an intermediary mechanism where the server acts as a predictive intermediary, calculating and sending multiple future predictions ahead of time. This intermediary layer absorbs the latency variability by pre-computing commands that the client can execute without real-time server communication.
2Reliability
If multiple predictions with sensor signal hypotheses are generated and transmitted, then latency compensation is achieved, but network bandwidth consumption increases
Solution Approach 1:
Instead of sending complete sensor data and recalculating everything at the server, the patent sends only differential updates and corrections from client to server. The server sends predictions, and the client sends back only the differences or corrections needed, reducing network traffic while maintaining prediction accuracy.
Solution Approach 2:
The server creates simplified copies or representations of sensor signals (sensor signal hypotheses) rather than processing complete sensor data streams. These hypothesis copies are sufficient for generating operation commands but require less bandwidth to transmit and process.
3Speed
If the client device performs local computational tasks, then real-time response is maintained, but hardware requirements and complexity increase
Solution Approach 1:
The client device is empowered to autonomously select predictions and generate operation commands based on received sensor signals and pre-sent predictions from the server. The client performs self-service by matching incoming sensor data against predicted hypotheses and executing commands without requiring complex local AI/ML infrastructure.
Solution Approach 2:
The patent replaces complex local computational mechanics (AI/ML inference engines, large local models) with a simpler information processing approach. Instead of heavy local computation, the client uses lightweight hypothesis matching and selection based on pre-computed predictions from the server.
Data Source
AI summary
The invention relates, amongst other things, to a method for operating a client device. If a first operation command is not received from a server device within a predetermined period of time in response to sending at least one first sensor signal, one of a plurality of predictions is selected depending on at least one first sensor signal and/or at least one currently detected sensor signal and a plurality of sensor signal hypotheses and the client device is operated with at least one operation command hypothesis of the selected prediction. The invention also relates, amongst other things, to a method for operating a server device.