logo-header

Percepta – A Lightweight Data Stream Processing System for Safe and Efficient Edge AI

As Edge AI continues to gain traction across industrial and distributed environments, the need for efficient, real-time data processing at the edge becomes increasingly important. Handling data streams close to the source enables faster decision-making, reduced latency, and improved system responsiveness.

In this context, Luis Lino Ferreira (ISEP) presents Percepta, a lightweight data stream processing system designed to support safe and efficient Edge AI applications.

The Need for Edge Data Stream Processing

Modern systems generate large volumes of data from sensors, devices, and connected services. Processing this data centrally can introduce delays, increase bandwidth usage, and limit scalability.

Edge computing addresses these challenges by moving computation closer to the data source. However, this shift also requires solutions that are:

  • Lightweight and resource-efficient
  • Capable of handling continuous data streams
  • Suitable for distributed and heterogeneous environments

Percepta is designed to meet these requirements by enabling efficient data processing directly at the edge.

What is Percepta?

Percepta is a lightweight data stream processing system that supports real-time handling of data flows in edge environments. It is designed to operate efficiently on constrained devices while maintaining scalability across distributed systems.

The system focuses on:

  • Processing streaming data in real time
  • Supporting safe and efficient execution of Edge AI workflows
  • Enabling scalable deployments across multiple edge nodes

Supporting Safe and Efficient Edge AI

One of the key challenges in Edge AI is balancing performance with system constraints such as limited computing resources and network capacity.

Percepta addresses this by:

  • Minimizing resource consumption while processing data streams
  • Enabling faster local decision-making
  • Reducing dependency on centralized infrastructure

This makes it suitable for applications where responsiveness and efficiency are critical.

Applicability in Distributed Systems

Percepta is particularly relevant for:

  • Industrial IoT environments
  • Cyber-physical systems
  • Distributed AI applications
  • Real-time monitoring and control scenarios

By supporting data processing at the edge, it contributes to more resilient and scalable system architectures.

Watch the Full Presentation

To learn more about Percepta and its role in Edge AI, watch the full presentation by Luis Lino Ferreira:
👉 https://youtu.be/iQWSaLb4jWk

Acknowledgment

This development is part of the Eclipse Arrowhead framework and the Arrowhead fPVN project (Grant Agreement No. 101111977), and has been supported by the AIMS5.0 project. This project has received funding from the Chips Joint Undertaking (Chips JU) and its members, including top-up funding by Finland, Denmark, Sweden, Spain, Italy, Romania, Portugal, Hungary, and France.