Skip to main content


Call for Papers: Distributed Signal Processing and Machine Learning for Heterogeneous Edge Computing

The capabilities of microcontrollers is been increasing over the past decades in terms of computation, storage and power consumption. Today, more than ever, we are able to deploy a large number of IoT devices with diverse computational and communication capabilities that are capable of carrying out a large number of heterogeneous task related to the processing and analysis of data while at the same time maintaining a lower-power consumption profile. The aforementioned device and task heterogeneity usually leads to major differences in computation and communication latency of serving IoT tasks with unpredictable dynamics due to human-in-the-loop. These constraints introduce critical challenges in network design and management for IoT. However, existing Internet of Things deployments still follow a cloud-centric architecture where data are forwarded to cloud services for analysis and storage, thus leaving the capabilities of the IoT devices almost completely underutilized. 

Recently the Edge computing approach was proposed as an attempt to shift computational and storage of tasks from being completely cloud-based to also involve computational elements residing on the edges of the network. First results indicate that we can move the analysis and interpretation of a wide number of sensor data traces within the edge devices and provide actionable alerts in an energy-optimized way. Executing simple motion analysis, sound classification, image recognition, gesture detection schemes on the edge devices, not only seems to remain a challenge, but becomes more essential in dynamic fog networks with time varying user demands. Adaptive Signal processing and online machine/deep learning algorithms can be combined for the robust identification of  complex events in non-stationary dynamic environments utilizing the available computational resources of the edge devices.

The aim of this Special Issue is to bring together researchers and practitioners from diverse fields of science and engineering working towards the realization of edge-computing based signal processing and machine/deep learning. We invite authors to submit original research articles, surveys and viewpoint articles related to recent advances and related technologies. We are particularly interested in presenting emerging technologies related to intelligent sensing utilizing edge resources that may have a significant impact on the field for years to come. We are open to papers addressing a broad range of topics, from foundational topics regarding the operating principles, and novel design principles for building future intelligent sensing; to papers presenting advanced frameworks and technological platforms; to pilots reporting innovative real-world deployments.

Topics of interest for the Special Issue include (but are not limited to):

  • Methodologies for studying, analyzing and building systems utilizing edge computing
  • Novel application scenarios, challenges and opportunities
  • Development challenges and approaches for intelligent sensing utilizing edge-computing
  • Programming frameworks for edge computing
  • Middle ware for edge computing
  • System software for edge computing and low-power devices
  • Data management and knowledge extraction in edge computing
  • Signal processing and Machine/Deep learning in resource-constraint devices
  • Real-time Signal processing and Machine learning
  • Distributed and Cooperative Signal Processing 
  • Heterogeneous Online Learning for Fog Computing
  • Prototype hardware for delivering intelligent sensing utilizing edge-computing
  • Hardware assisted Machine learning
  • Energy Efficient Signal processing and Machine learning
  • Real-world deployments; pilots of intelligent sensing utilizing edge-computing
  • Evaluation of intelligent sensing utilizing edge-computing

Deadline for Submissions:

28th February 2020

Lead Guest Editor:

Ioannis Chatzigiannakis, Sapienza University of Rome, Italy

Guest Editors:

  • Aris S. Lalos, Athena Research Institute, Greece
  • Christos Kotselidis, University of Manchester, UK
  • Polychronis Xekalakis, Esperando Technologies, USA