PARADICE

As countries combat the COVID-19 pandemic, they realise more that not only health sector is struggling, agriculture is also facing real challenges. Threats are serious: closing borders paralyzes imports/exports, the lockdown participated in the disruptions in the food demand and therefore supply chains, etc. Together with the danger of the climate change, these threats confirm that the agriculture field is a national priority and a clue to preserve sovereignty. With precision agriculture in mind, our project focuses on providing a complete architecture, centered on the use of networked Unmanned Aerial Vehicles (UAVs known also as drones), and able to run efficiently artificial intelligence algorithms.

Drones are acting as a homogeneous group to achieve a common mission within the monitored area. To reach that goal and in order to coordinate their actions, they communicate with each other and with terrestrial devices (IoT, base stations, etc). For any embedded application, drones have the following fundamental actions: sensing, data collection and processing and mobile data relaying. Drones behave smartly (run machine learning algorithms to plan paths and detect abnormal situations). 

The challenges addressed by the project are related to technology: integrate building blocks (such as UAVs, base stations, sensors), communications, embedded and edge machine learning. We found that the same identical architecture can address various scenarios such as: 

• Plant and soil health: early detection of plants pandemics, 

• Smart irrigation and fertilization, and 

• Monitor the cattle (detection of pandemics and cattle theft).

For this project, we intend to develop a prototype for early plants pandemics detection. Using near-infrared, drones are able to identify stress in a plant 10 days before it becomes visible to the eye.  We will use the artificial intelligence techniques as they are excellent solutions for deciding in complex environments during system runtime (drone navigation, early pandemics detection, etc). At the same time, they require efficient hardware support for machine learning (ML). We focus on both UAVs radio communications (Air-to-Air and Air-to-Ground communications) as they are the cornerstone to ensure a coherent UAVs group navigation. The underlying technology which makes the project possible will be our primary focus (TinyML, embedded ML, edge ML, WLAN IEEE 802.11, 4G/5G cellular networks, IoT communications, etc).