Home automation risk assessment and mitigation via machine learning

Patent metadata

Field	Value
Publication number	US-11477230-B2
Application number	US-202016819242-A
Country	US
Kind code	B2
Filing date	Mar 16, 2020
Priority date	Mar 16, 2020
Publication date	Oct 18, 2022
Grant date	Oct 18, 2022

Abstract

An approach for identifying mitigation solution based on critical situations is disclosed. The approach includes detecting one or more critical situations associated within a structure and detecting one or more location of one or more users in the structure. The approach retrieves a user-knowledge corpus based on one or more smart IoT devices or from existing database. Furthermore, the approach retrieves a critical situation knowledge corpus from various servers and creates risk mitigation action plans to address the one or more critical situations. The approach selects an optimal plan, by leveraging machine learning through combinatorial optimization technique, from the existing risk mitigation action plans and executing the optimal plan.

First claim

What is claimed is:

1. A computer-implemented method for identifying mitigation solution based on critical situations, the computer-implemented method comprising:
detecting one or more critical situations associated within a structure;
detecting one or more location of one or more users in the structure;
retrieving a user-knowledge corpus based on one or more smart IoT devices from a first database;
retrieving a critical situation knowledge corpus from a second database;
creating one or more risk mitigation action plans to address the one or more critical situations; wherein creating one or more risk mitigation action plan based on the one or more critical situations comprises:
analyzing the user-knowledge corpus and the critical situation knowledge corpus related to the one or more critical situations;
creating one or more risk mitigation action plans based on the analyzed user knowledge corpus and critical situation knowledge corpus;
assigning weighted risk scores to the one or more risk mitigation action plans;
selecting an optimal plan, by leveraging machine learning through combinatorial optimization technique, from the one or more risk mitigation action plans; and executing the optimal plan.

2. The computer implemented method of claim 1 , wherein detecting one or more critical situations within the structure comprises:
receiving sensor data from the one or more IoT sensors; and
receiving warning data from the one or more knowledge servers.

3. The computer implemented method of claim 1 , wherein retrieving the user-knowledge corpus based on the one or more smart IoT devices from the first database comprises:
identifying a usage pattern one or more users in the structure based on the one or more smart IoT devices, wherein the usage pattern comprises of one or more activities and wherein the one or more activities comprises of a type and sequence and wherein the one or more smart IoT devices comprises a location of the one or more IoT devices and dependency amongst the one or more IoT devices.

4. The computer implemented method of claim 1 , wherein retrieving the critical situation knowledge corpus separate from the one or more servers from the second database comprises:
connecting to the one or more servers;
querying the one or more servers for data related to the one or more critical situations;
responsive to an empty query of the critical situation knowledge corpus, creating a new critical situation knowledge corpus; and
downloading the data, wherein the data comprise of information from IoT devices, camera feeds, investigation reports, evacuation plans, rescue operation steps and guidelines for one or more types of critical situations.

5. The computer implemented method of claim 1 , wherein selecting an optimal plan, by leveraging machine learning through the use of combinatorial optimization technique, from the one or more risk mitigation action plans comprises: determining whether the weighted score of the one or more risk mitigation action plan exceeds a risk threshold; and responsive to determining that the one or more risk mitigation action plan does not exceed the risk threshold, selecting the one or more risk mitigation action plan as the optimal plan.

6. The computer implemented method of claim 1 , wherein executing the optimal plan comprises:
transmitting to an output component to execute the optimal plan, wherein the output component communicates to one or more action devices comprises of IoT lamps, HVAC, home automation/security system, PC, mobile devices and speakers.

7. A computer program product for identifying mitigation solution based on critical situations, the computer program product comprising:
one or more computer readable storage media and program instructions stored on the one or more computer readable storage media, the program instructions comprising:
program instructions to detect one or more critical situations associated within a structure;
program instructions to detect one or more location of one or more users in the structure;
program instructions to retrieving a user-knowledge corpus based on one or more smart IoT devices from a first database;
program instructions to retrieve a critical situation knowledge corpus from a second database;
program instructions to create one or more risk mitigation action plans to address the one or more critical situations;
wherein program instructions to create one or more risk mitigation action plan based on the one or more critical situations comprises:
program instructions to analyze the user-knowledge corpus and the critical situation knowledge corpus related to the one or more critical situations;
program instructions to create one or more risk mitigation action plans based on the analyzed user knowledge corpus and critical situation knowledge corpus; and
program instructions to assigning weighted risk scores to the one or more risk mitigation action plans;
program instructions to select an optimal plan, by leveraging machine learning through combinatorial optimization technique, from the one or more risk mitigation action plans; and program instructions to execute the optimal plan.

8. The computer program product of claim 7 , wherein program instructions to detect one or more critical situations within the structure comprises:
program instructions to receive sensor data from the one or more IoT sensors; and
program instructions to receive warning data from the one or more knowledge servers.

9. The computer program product of claim 7 , wherein program instructions to retrieve the user-knowledge corpus based on the one or more smart IoT devices from the first database comprises:
program instructions to identify a usage pattern one or more users in the structure based on the one or more smart IoT devices, wherein the usage pattern comprises of one or more activities and wherein the one or more activities comprises of a type and sequence and wherein the one or more smart IoT devices comprises a location of the one or more IoT devices and dependency amongst the one or more IoT devices.

10. The computer program product of claim 7 , wherein program instructions to retrieve the critical situation knowledge corpus separate from the one or more servers from the second database comprises:
program instructions to connect to the one or more servers;
program instructions to query the one or more servers for data related to the one or more critical situations;
responsive to an empty query of the critical situation knowledge corpus, program instructions to create a new critical situation knowledge corpus; and
program instructions to download the data, wherein the data comprise of information from IoT devices, camera feeds, investigation reports, evacuation plans, rescue operation steps and guidelines for one or more types of critical situations.

11. The computer program product of claim 7 , wherein program instructions to select an optimal plan, by leveraging machine learning through the use of combinatorial optimization technique, from the one or more risk mitigation action plans comprises:
program instructions to determine whether the weighted score of the one or more risk mitigation action plan exceeds a risk threshold; and
responsive to program instructions to determine that the one or more risk mitigation action plan does not exceed the risk threshold, program instructions to select the one or more risk mitigation action plan as the optimal plan.

12. The computer program product of claim 7 , wherein program instructions to execute the optimal plan comprises:
program instructions to transmit to an output component to execute the optimal plan, wherein the outpu

Assignees

• IBM

Inventors

• Baughman Aaron K
• Nagar Raghuveer Prasad
• Kwatra Shikhar
• Rakshit Sarbajit K

Classifications

• G06N5/01 Primary

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06N20/00

  Machine learning · CPC title

• G16Y10/80

  Homes; Buildings · CPC title

• G06N5/04

  Inference or reasoning models · CPC title

• G16Y30/10

  Security thereof · CPC title