Back to results
US 10,762,616 B2Granted
Method and system of analytics system balancing lead time and accuracy of edge analytics modules
AssigneeHitachi, Ltd.
InventorDaisuke Ito
Life of this patent
Estimating the term…
Abstract
Systems and methods described herein are directed to changing models and analytics algorithms on an edge node from a core server, which can be useful in situations such as optimized factories when the edge node is physically close to data sources such as within a factory plant. The core server or cloud runs analytics algorithms and models concurrently with data received from edge nodes, and replaces the edge node with more accurate analytics algorithms and models as applicable.
Claims15
An apparatus configured to manage a plurality of edge nodes, the apparatus comprising:
- a processor, configured to:
- for data received from an edge node from the plurality of edge nodes, process the data using a first analytics model that is utilized by the edge node,
- process the data using a second analytics model, the second analytics model trained through an analytics trainer managed at the apparatus;
- and
- for the second analytics model determined to have better accuracy than the first analytics model, update the edge node with the second analytics model;
- wherein the apparatus further comprises a storage configured to store a plurality of analytics algorithms, wherein the processor is further configured to:
- execute the plurality of analytics algorithms on the data;
- determine a first one of the plurality of analytics algorithms having a highest accuracy among the plurality of analytics algorithms;
- and
- for the first one of the plurality of analytics algorithms having higher accuracy than a second one of the plurality of analytics algorithms that is executed in the edge node, update the edge node with the one of the plurality of analytics algorithms.
7 more elements
A method for managing a plurality of edge nodes, comprising:
- for data received from an edge node from the plurality of edge nodes, processing the data using a first analytics model that is utilized by the edge node,
- processing the data using a second analytics model, the second analytics model trained through an analytics trainer;
- and
- for the second analytics model determined to have better accuracy than the first analytics model, updating the edge node with the second analytics model;
- managing a plurality of analytics algorithms;
- executing the plurality of analytics algorithms on the data;
- determining a first one of the plurality of analytics algorithms having a highest accuracy among the plurality of analytics algorithms;
- and
- for the first one of the plurality of analytics algorithms having higher accuracy than a second one of the plurality of analytics algorithms that is executed in the edge node, updating the edge node with the one of the plurality of analytics algorithms.
6 more elements
A system, comprising:
- a plurality of edge nodes;
- and
- an apparatus comprising:
- a processor, configured to:
- for data received from an edge node from the plurality of edge nodes, process the data using a first analytics model that is utilized by the edge node,
- process the data using a second analytics model, the second analytics model trained through an analytics trainer managed at the apparatus;
- and
- for the second analytics model determined to have better accuracy than the first analytics model, update the edge node with the second analytics model;
- wherein the apparatus further comprises a storage configured to store a plurality of analytics algorithms, wherein the processor is further configured to:
- execute the plurality of analytics algorithms on the data;
- determine a first one of the plurality of analytics algorithms having a highest accuracy among the plurality of analytics algorithms;
- and
- for the first one of the plurality of analytics algorithms having higher accuracy than a second one of the plurality of analytics algorithms that is executed in the edge node, update the edge node with the one of the plurality of analytics algorithms.
10 more elements
Dependencies are inferred from the claim text.