{"id":1292,"date":"2026-04-13T22:28:01","date_gmt":"2026-04-13T21:28:01","guid":{"rendered":"https:\/\/www2.isep.ipp.pt\/softcps\/?p=1292"},"modified":"2026-04-16T14:01:22","modified_gmt":"2026-04-16T13:01:22","slug":"new-framework-improves-real-time-data-processing-for-ai-at-the-edge","status":"publish","type":"post","link":"https:\/\/www2.isep.ipp.pt\/softcps\/?p=1292","title":{"rendered":"New framework improves real-time data processing for AI at the edge\u00a0"},"content":{"rendered":"\n

SoftCPS Researchers contributed to the development of Percepta, a lightweight data stream processing framework designed to efficiently manage real-time data streams for artificial intelligence applications operating at the edge. <\/p>\n\n\n\n

Artificial intelligence systems increasingly rely on continuous data streams generated by connectedIoT devices. From smart homes to industrial environments, sensors and digital platforms produce large amounts of information that must be processedin real-time  to support automated decision-making. <\/p>\n\n\n\n

However, managing these data streams is not straightforward. Data often comes from multiple providers using different communication protocols, formats and update frequencies. Processing all this information in the cloud may also introduce latency and increase bandwidth usage, which can limit the responsiveness of AI systems. <\/p>\n\n\n\n

To address these challenges, the team  developed Percepta. It acts as an intermediary layer between data sources and Machine Learning models. <\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Its main goal is to collect information from multiple providers, translate it into a consistent format and prepare it for continuous processing by AI algorithms. <\/p>\n\n\n\n

The framework follows a modular architecture that allows it to adapt to different use cases and data sources. Incoming data is first handled by dedicated modules capable of managing the communication protocols used by each provider. The information is then translated into a standardized format so that it can be processed consistently regardless of its origin. <\/p>\n\n\n\n

Once standardized, the system performs several processing operations to improve the structure and reliability of the incoming streams. These include handling missing values, removing duplicate entries, correcting out-of-order data and harmonising data generated at different frequencies. After this step, the processed data is delivered to the Machine Learning model responsible for generating predictions or decisions for missing data. <\/p>\n\n\n\n

Percepta also stores the processed information and model outputs, enabling further analysis and supporting the retraining of AI models. <\/p>\n\n\n\n

The framework was initially developed within the OPEVA (Optimization of Electric Vehicle Autonomy) project to support EnergAIze, an artificial intelligence model designed to optimise household energy consumption, and later expanded in the context of Arrowhead fPVN project and Arrowhead framework. EnergAIze analyses real-time energy data to make decisions such as charging electric vehicles when photovoltaic production is high or activating appliances when renewable energy is available , among other more complex functionalities. <\/p>\n\n\n\n

To operate effectively, the model relies on continuous data streams such as photovoltaic production levels, electricity consumption metrics and electric vehicle flexibility information. Precept helps manage these heterogeneous data sources by collecting, harmonising and preparing the data before it reaches the AI model. <\/p>\n\n\n\n

The framework can be applied to many other scenarios that require reliable real-time data processing for AI systems, including Internet of Things environments, smart infrastructures and industrial automation. <\/p>\n","protected":false},"excerpt":{"rendered":"

SoftCPS Researchers contributed to the development of Percepta, a lightweight data stream processing framework designed to efficiently manage real-time data streams for artificial intelligence applications operating at the edge.  Artificial intelligence systems increasingly rely on continuous data streams generated by connectedIoT devices. From smart homes to industrial environments, sensors and digital platforms produce large amounts of information that must be processedin […]<\/p>\n","protected":false},"author":8,"featured_media":1295,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[5],"tags":[],"class_list":["post-1292","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"blocksy_meta":"","_links":{"self":[{"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=\/wp\/v2\/posts\/1292","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1292"}],"version-history":[{"count":2,"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=\/wp\/v2\/posts\/1292\/revisions"}],"predecessor-version":[{"id":1297,"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=\/wp\/v2\/posts\/1292\/revisions\/1297"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=\/wp\/v2\/media\/1295"}],"wp:attachment":[{"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1292"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1292"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www2.isep.ipp.pt\/softcps\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1292"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}