Big Data & Analytics

LET DATA TAKE DECISIONS

THE SMART WAY

AUTOMATED IMPROVEMENT THROUGH EXPERIENCE

X-BRAiN provides a wide set of AI and machine learning based services that can be used in dedicated application scenarios.

Detection & Analsys

DESCRIPTIVE ANALYSIS

Statistical Description of Sensor Data and Events

X-BRAiN provides several dashboard options for getting statistical overviews of sensor measurements and associated events. This includes not only the embedded X-BRAiN dashboard where comprehensive graphs and data tables are available, but also dashboard options that are integrated with third party services like Google Data Studio, Amazon CloudWatch and more.

Anomaly Detection (Correlation Between Data & Events)

Heat maps for visualising the data correlation between all sensor measurements and associated events are available within X-BRAiN. This ensures that anomalies with respect to the sensor measurements can be detected, especially when correlations are not obvious to the human eye.

Temporal and Seasonal Analytics

X-BRAiN provides an overview and comparison of seasonal and temporal sensor measurements. A dedicated section with graphs and data tables is available where sensor data for each season or other time periods are given. This includes the comparison of the associated data with former seasons in years past.

Impact Analysis

When gathering and managing large amounts of data, correlations will usually be present, causing an impact in case the associated data is changing. X-BRAiN is providing capabilities for highlighting those impacts that are described as events.

Operational Condition of the Site

Depending on the application scenario, it is necessary to be aware of the operational condition of a site. This applies to domains like Smart Agriculture or Smart Aquariums, where X-BRAiN is able to detect if plants are entering their growing state in order to trigger associated actions like adjusting the watering and fertilising behaviour. For detecting the operational state, multiple reference objects will be added to the X-BRAiN inventory along with reference data, where ML capabilities will be applied.

Machine Learning

PATTERN RECOGNITION

Complex event Detection Strategies

X-BRAiN is able to detect complex events by applying and combining several machine learning techniques. This includes the chaining of independent events, the involvement of past events as well as factoring in the associated measurements. This allows to establish custom detection strategies for custom events within specific application scenarios.

Recognition of Event Sequences

Event sequences can heavily impact ecosystems that are monitored by IoT devices. Therefore, it is essential to detect those sequences as soon as they appear. This includes the detection of weather extremes in Smart Agriculture, like heavily alternating temperatures during several days, that can threaten the condition of sensitive plants.

Detection of Drift Sequences

X-BRAiN provides comprehensive abilities for predictive maintenance, which includes the detection of drift sequences. Applied to the IoT devices that are installed in the field, broken and inaccurate devices will be detected as soon as a drift in their measurements occurs, especially in relation to reference devices that are mounted nearby.

Monitoring & Forecasting

PREDICTIVE ANALYSIS

Predictive Health Monitoring

X-BRAiN is able to perform predictive health monitoring that is tailored to the applied application scenario. In this context, alerts and warnings will be raised in case a potential threat has been forecasted. By using big data capabilities that are linked to our internal databases and all inherent data sets, accurate forecasting can be achieved by considering related seasonal and temporal data as well as patterns, including external services like weather forecasts.

Root Cause Analysis

Although X-BRAiN is constantly performing several AI and machine learning analysis, there are always events that can not be detected or prevented. In order to identify the associated root cause and consider the overall behaviour in future scenarios, refined machine learning techniques can be applied to highlight smallest anomalies and drifts compared to the usual data behaviour.

Multivariate Time Series Forecasting

In order to forecast the measurements of installed IoT devices, multivariate time series forecasting is applied by X-BRAiN. This machine learning technique benefits from our large set of history data that is available in our internal databases, both per device type and application scenario.

Alarms & Training

PRESCRIPTIVE ANALYSIS

React on Issues and Alarms by Automatic Recommendations

X-BRAiN is able to react to issues and anomalies that have been detected. This can happen automatically by triggering associated actuators, or indirectly by defining recommendations for users. Those recommendations are raised based on historised data and events, or can be defined manually by the user.

Training Data for Machine Learning Techniques

X-BRAiN provides a comprehensive set of data for several application scenarios in the IoT domain. This data is highly valuable as it acts as a fundament for applying several forecasting techniques. Those data sets can be used as a dedicated service in order to train neural networks, ensuring high forecasting accuracies.