KnowDetector

This repository is dedicated to sharing the tool KnowDetector and the data for the ASE 2023 paper entitled 'Detecting Smart Home Automation Application Interferences with Domain Knowledge'.

Introduction

The KnowDetector parses a set of automation applications (group data), extracting their conditions, events, and actions. It forms a basic Event-Condition-Action (ECA) network based on the control relationships. Then, using a knowledge graph, it enriches the semantic information to form an automated semantic network. Finally, pattern matching is performed based on the proposed sub-graph patterns of interference, in conjunction with semantic information, to detect interference.

Getting started

Environment requirements

• Java 1.8

The dependencies can be found in pom.xml

How to run

• Quick run

1. We provide a JAR package (knowdetector_jar/knowdetector.jar) that can be executed using the following instructions.

2. Prior to executing the code, please make sure that you have placed the grouped test files into the knowdetector_jar/testFiles folder. We have already provided some group data (test data) in the knowdetector_jar/testFiles directory.

3. java -jar knowdetector.jar --testFilePath --resultFilePath

   For example: java -jar knowdetector.jar ./testFiles/10Loops ./result

4. The results will be stored in knowdetector_jar/result

The results of the aforementioned instructions are as follows:

The results are saved in three columns, indicating the corresponding interference pattern, the involved files, and the related edges and vertices.

• Build from source code

1. Configure the programs using your IDE, such as IntelliJ IDEA.
2. Prior to executing the code, please make sure that you have placed the grouped test files into the src/main/resources/testFiles folder. We have already provided some group data (test data) in the src/main/resources/testFiles directory.
3. Run the file src/main/java/Detection.java to detect automation interferences.
4. The results will be stored in the file src/main/resource/results/result.csv

Knowledge graph data

The knowledge graph data is located in the /data directory. For the sake of simplicity, we have presented the knowledge graph data using .csv files. Specifically:

• DeviceModel.csv captures the device models.

• relation.csv captures the relations between device services.

Detailed description

The DeviceModel.csv contains 17 columns.

• Id. Unique ID used for counting
• device. Device category, such as AC (Air Conditioner), light.
• deviceName. Specific device names, such as OPPLELamp and yeelightLEDColor.
• devType. Device types extracted from device metadata.
• prodId. Product Ids extracted from device metadata.
• capabilityId. Unique ID of device capabilities.
• capabilityDescription. The capability description.
• commandId. Unique ID of commands.
• valueType. Parameter types of commands, including "enumeration" types and "range" types.
• enum_values. Detailed enum. data.
• value. The specific value of the enum data.
• max, min. The range value.
• serviceId. Unique ID for device service, it is composed of device-deviceName_capabilityId_commandId_value.
• serviceDescription. The service description.
• influencedChannels. Physical channels affected by the device service.
• influenceType. Types of impact, including "hasDecreased" and "hasIncreased".

The relation.csv contains 4 columns.

• Id. Unique ID used for counting
• sourceService. Source service.
• targetService. Target service.
• relation. Relationships between device services, including "implied" and "exclusive".

Publicly shared data

We provide two set of data, i.e., IFTTT data and the commercial data, in directories data/IFTTT and data/commercial data, respectively.

IFTTT data

The raw data is listed in IFTTT_raw_data.json. We have transformed the data into a JSON format that is easy to parse, based on the provided knowledge model. You can find the file in the IFTTT_data folder.

Commercial data

We converted and anonymized a portion of commercial automation data in the form of annotated JSON files. Commercial data is located in the data/commercial data directory, which includes all the data (under the all_data directory) as well as the grouped data (under the group_data directory, each group is a set of automations). The group_data can be used as a benchmark for testing. The JSON file is as follows:

'''

{"id": "a753cd54-f010-4328-8c69-22a0d38bd1a1",
"trigger": {
    "conditions": [
        {
            "condId": "cond.temperature",
            "physical": "temperature",
            "params": {
                "operator": ">",
                "value": {
                    "range":"30"
                }
            }
        }
    ],
    "events": [
        {
            "devType": "00A",
            "deviceId": "6b3cdd3a-5fa0-435f-a5ef-254783a46525",
            "eventId": "eff4cefc-9dc8-46c0-8e55-7bec9c0c82fe",
            "params": {
                "capabilityId": "audioPlayState",
                "command": "pause",
                "value": "1"
            },
            "prodId": "001T"
        }
    ]
},
"actions": [
    {
        "actions": [
            {
                "actionId": "469fbd2d-ad13-4cae-9486-c8b5a5e7056a",
                "devType": "00A",
                "deviceId": "6b3cdd3a-5fa0-435f-a5ef-254783a46525",
                "params": {
                    "capabilityId": "playMusic",
                    "command": "play",
                    "value": "aiting:5373510"
                },
                "prodId": "001T"
            },
            {
                "actionId": "fbf0e4aa-dd96-4930-ac04-29100ceb3c98",
                "devType": "00A",
                "deviceId": "6b3cdd3a-5fa0-435f-a5ef-254783a46525",
                "params": {
                    "capabilityId": "audioPlayState",
                    "command": "pause",
                    "value": "1"
                },
                "prodId": "001T"
            }
        ],
        "delay": {
            "delaySync": false
        }
    }
]}

'''

Publication

If you are interested in our work, you can find more details in our paper listed below. If you use our dateset and tool, please cite our paper.

Detecting Smart Home Automation Application Interferences with Domain Knowledge

38th International Conference on Automated Software Engineering (ASE'23)