Data-Driven Embedded/IoT Design
A Half-hour Webinar and Live Demo
In modern embedded systems, the volume and criticality of sensor data is increasing. The concept of “Sensor Fusion” depicts the way in which we must integrate and respond to this data on the fly. Your data architecture and real-time flow must be an early consideration when contemplating embedded or IoT product design, not shoehorned after you’ve chosen the software and hardware – this is the essence of “Data Driven Development”
Traditionally, structuring real-time streaming data as a database doesn’t always come naturally when designing a resource constrained system, but increasing data bandwidth and value means we should review this orthodoxy to avoid unnecessary compromises. We will demonstrate how you can create flexible, configurable data structures which allow data to drive your system logic in real-time – combining, downsampling and reacting intelligently to the data stream locally rather than simply recording, uploading and processing offline or in the cloud.
In systems with a distributed sensor/controller topology we will demonstrate how it’s possible to seamlessly query data across connected nodes, including low-power microcontrollers and high-performance microprocessors – allowing the live data analysis to span the network.
Where machine learning is to be employed we need an approach that allows edge data to be pre-processed and conditioned in real-time. Where AI is used for example for anomaly detection, data capture can be dynamically adjusted according to AI outputs, empowering predictive maintenance, and other modalities where we gain value in real-time from data despite high bandwidth and/or complexity.
All this can be accomplished in a cybersecurity standards-ready and functional safety capable software environment.
