Dynamic AI Gateway for Neural Network Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing approaches to autonomous driving systems face inefficiencies due to mismatched processing resources between central servers and edge servers, leading to excessive data traffic and battery consumption in autonomous vehicles, as they rely on uploading large sensor data to remote clouds for processing.
Innovation Solution
A dynamic AI gateway system that monitors and redistributes artificial neural network (ANN) neurons across multiple processing nodes based on real-time load monitoring, enabling efficient resource utilization and reducing data traffic by processing data closer to the source.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If large sensor data is uploaded to remote clouds for processing, then processing capability is improved, but data traffic and battery consumption increase excessively
Solution Approach 1:
The patent segments the ANN into multiple neurons distributed across different processing nodes (edge devices and central server). This allows processing to be divided and executed locally at the edge rather than requiring all data to be uploaded to the cloud, thereby reducing data traffic and battery consumption while maintaining processing capability.
Solution Approach 2:
The patent implements local processing of sensor data at the edge device by distributing ANN neurons to process data locally before only necessary results are transmitted to the central server. This local quality approach reduces the amount of data needing to be uploaded, thus reducing energy consumption and data traffic.
2Power
If processing resources are centralized at the cloud server, then processing power is concentrated, but resource utilization becomes mismatched and inefficient
Solution Approach 1:
The patent segments the centralized processing power across multiple nodes by distributing ANN neurons to both edge devices and the central server. This segmentation allows each node to process data independently based on its capabilities and workload, improving overall resource utilization efficiency while maintaining concentrated processing power where needed.
Solution Approach 2:
The patent implements dynamic redistribution of ANN neurons across processing nodes based on real-time load monitoring. This dynamic approach allows the system to adapt resource allocation to current conditions, optimizing resource utilization efficiency by moving processing tasks to nodes with available capacity rather than relying on fixed centralized architecture.
3Loss of energy
If all ANN processing is performed at the edge device, then data traffic is minimized, but processing resources may be overloaded
Solution Approach 1:
The patent merges the processing capabilities of edge devices and central server into a hybrid distributed system. By combining local edge processing with centralized cloud processing, the system minimizes data traffic by processing data locally while maintaining processing stability through the shared central server that can offload tasks when edge resources become overloaded.
Solution Approach 2:
The patent implements feedback mechanisms through load monitoring that track the processing status of neurons at each node. This feedback allows the system to dynamically adjust the distribution of ANN neurons based on real-time conditions, ensuring processing stability by redistributing tasks away from overloaded nodes while maintaining minimal data traffic through the established edge-processing architecture.
Data Source
AI summary
Methods, systems, and apparatus related to dynamic distribution of an artificial neural network among multiple processing nodes based on real-time monitoring of a processing load on each node. In one approach, a server acts as an intelligent artificial intelligence (AI) gateway. The server receives data regarding a respective operating status for each of monitored processing devices. The monitored processing devices perform processing for an artificial neural network (ANN). The monitored processing devices each perform processing for a portion of the neurons in the ANN. The portions are distributed in response to monitoring the processing load on each processing device (e.g., to better utilize processing power across all of the processing devices).