Pushing AI at the Edge: Intelligent and Contextually Aware Sensor Networks in Harsh or Secluded Environments

Pushing AI at the Edge: Intelligent and Contextually Aware Sensor Networks in Harsh or Secluded Environments

As the number of IoT- connected devices is projected to increase to 43 billion by 2023, there is a common expectation that the vast majority of these will become (artificially) intelligent. Intelligence-enabled edge computing and AI on edge are therefore facing the same surging demand in integrated capabilities on the device side as IoT was, in terms of connectivity, only several years ago. Next-generation solutions rely on a new suite of features to ensure the commercial success of the new intelligent “things”. Shrinking in size dedicated hardware is becoming more powerful, upgraded with new standards (such as COM-HPC), hence expanding the possibilities for heavier machine learning workloads at the edge. A growing number of smaller and more specialized sensors are ubiquitously embedded into everything, fostering a new trend: real-time locally made educated decisions in response to the information acquired straight from the surrounding environment, precluding interactions with the cloud.

Contextual awareness and recommender features will gradually complete the intelligent behavior of all connected “things” around us, providing augmented sensing experiences but what should we expect from intelligent and contextually aware sensor networks, especially in harsh or secluded environments? The answer is straightforward: A proactive and intelligent response based on their awareness of both the ambient environment and the contextual situation, even when withstanding strong winds, floods, extreme temperatures or mechanical shocks.

This presentation will showcase the presenter’s inceptive work as part of the technical team of a conservation project focusing on endangered populations of giraffes and leopards in the context of surface roads infrastructure modernization and urban expansion.