ADLINK Technology has filed a patent for a query-oriented event recognition system. The system includes a capturing unit, an analyzing inference engine assembly, a synchronization processing unit, and an event definition unit. The system captures data flows with time points and data, analyzes the data using inference engines, generates inference data sets, and matches them with event query results. GlobalData’s report on ADLINK Technology gives a 360-degree view of the company including its patenting strategy. Buy the report here.
According to GlobalData’s company profile on ADLINK Technology, data center cooling systems was a key innovation area identified from patents. ADLINK Technology's grant share as of June 2023 was 1%. Grant share is based on the ratio of number of grants to total number of patents.
Query-oriented event recognition system
A recently filed patent (Publication Number: US20230196772A1) describes a query-oriented event recognition system. The system includes a capturing unit that captures data flows consisting of time points and corresponding data. An analysis and inference engine assembly, connected to the capturing unit, analyzes and reasons the data at each time point using multiple inference engines to generate inference data. A synchronization processing unit then generates an inference data set for each time point based on the data flows and inference data. Finally, an event definition unit uses a query statement to find an event query result that matches the inference data set.
The patent also mentions that the inference engines can be connected in series or parallel. Additionally, a judging and warning unit is included in the system to generate a warning signal based on the hazard level of the event query result. The event definition unit consists of a first and second event definition unit, where the first unit finds a first event query result matching the inference data set using a first query statement, and the second unit finds a second event query result matching the first event query result and the inference data using a second query statement.
Furthermore, the patent states that the inference data can be stored in metadata and a storage unit can be used to store the inference data. The query statement used in the system can be selected from graph theory, set theory, semantics, and logical judgment. The data being analyzed by the system includes image data and sound data.
In addition to the system, the patent also describes a query-oriented event recognition method. The method involves capturing data flows, analyzing and reasoning the data at each time point using multiple inference engines, generating inference data sets for each time point, and finding an event query result that matches the inference data set through a query statement. The method can also include a step of generating a warning signal based on the hazard level of the event query result. The event definition unit in the method follows the same process as described in the system, with a first and second event definition unit using different query statements to find matching event query results.