Meslova Sensor Data Processing Platform
Simple . Secure . Scalable
Get Valuable Insights including Pre-Failure and Failure Predictions, Notifications and Decision Support for discrete and IoT enabled Devices in a simple, scalable and secure manner. Whether you collect data using standards-based sensors, such as SCADA systems, Distributed Acoustic Sensing (DAS) equipment or proprietary hardware like many other organisations, the data would grow to terabytes or even petabytes in a short operational period.
One of the biggest challenges faced by organizations worldwide is not having efficient tools to process the large datasets, especially for sensor data which is known to have noise and is vulnerable to all kinds of sampling issues. Typically, organisations try to address this by either acquiring discrete proprietary applications for each sensor type or build their own application stack.
Proprietary sensor data processing applications address various sensor specific aspects but tend to be very endpoint specific and monolithic in nature. Building own application stack will solve business specific problems but require excessive amount of work in dealing with the standard data processing techniques which are mostly repetitive in nature.
Meslova Sensor Data Processing Platform brings in best of the both worlds and addresses the implementation issues in the best possible manner. With Out-of-the-box support for variety of sensor types involving multiple technologies, ability to process large datasets, built-in noise modelling, highly flexible configuration to indicate nature of different parameters and highly intuitive user interface, time required for customization and the total cost of ownership will be up to 70% lesser compared to operating discrete applications. Customers will get the best value and business outcomes from the Advanced Deep Learning Technology embedded in the platform.
Data Acquisition
-
Diverse Data Sources
Data acquisition from different sensor types with various characteristics
Support input in the form of human readable, standards-based as well as proprietary formats with ICD
Native support for multiple domains (Agriculture, Water, Oil & Gas and Defence)
Automatic conversion of binary input data to human readable formats including CSV, XLS, etc
-
Enormous Data Collection
Process volumes of data spread across number of files
Cached and Streaming Data
Supports Structured, Logs and Raw Data
-
Data Prioritization
Understands the data on the fly with domain specific awareness
Data Prioritization and Deduplication based on user configurable parameters
Processing
-
Scaling and Normalization
Normalize multi sensor and multi scale data
Binning the appropriate data
-
Manual and Automatic Feature Management
Parameter Selection
Correlation
Feature Extraction
Outlier Detection and Elimination
Noise Modelling and Consumption
Artificial Intelligence
-
Automatic Detection of different Data Characteristics
Density
Pattern
Time Series
Inter-dependency
-
Employs Deep Neural Network Architecture
DNN and CNN based Autoencoders
Generative Adversarial Networks
Supervised, Semi-supervised, Unsupervised and Reinforced Learning
Local and Global Optimization
Online and Incremental Learning
-
Learns complex data relationship
Correlation
Progression
Non-Linearity
Time Series
Geographic Awareness
-
Capabilities
User Interface
