An effective monitoring framework for biodiversity conservation that combines ECOacoustics with Artificial Intelligence and satellite Data transmission and elaboration

The global loss of biodiversity is one of the most pressing challenges of our time, calling for the rapid development of effective conservation protocols to protect, preserve, and restore the health and integrity of natural ecosystems. At the Italian level, there are several categories and levels of protected areas, but many of them struggle to achieve the required targets for nature conservation. A critical obstacle preventing the proper deployment of conservation protocols is the lack of an effective monitoring framework, which should guarantee fast, accurate, and efficient methods to conduct systematic environmental surveys. We thus urgently need tools for the rapid assessment of wildlife diversity and population dynamics at large scale and high spatiotemporal resolution, from individual animals to global densities. The aim of this project is to address this key issue by developing a scalable framework for the diffuse long-term management and monitoring of remote natural areas of biological relevance. The proposed approach will be based on a distributed network of smart sensors that continuously collect and process relevant environmental signals. Sensing nodes will be equipped with general sensing capabilities (e.g., temperature, humidity, solar radiation, etc.) as well as with high sensitivity microphones, which will be used to monitor ambient sounds through ecoacoustic analysis. The system will exploit advanced signal processing techniques stemming from recent advances in artificial intelligence and deep learning research, in order to automatically detect and analyse relevant information on the fly. A compressed summary of the information of interest will be then transmitted to the central operating station via satellite communication for further elaboration and interpretation. Such framework will allow for continuous and systematic monitoring of remote environments, both during generic survey operations and when specific target events are detected, causing a trigger alert. Our research approach will be highly interdisciplinary, and will capitalize on recent achievements in the fields of ecoacoustics, machine learning, and satellite telecommunication to effectively collect, analyse and catalyse data transfer; at the same time, it will further push the frontiers of research and technology by studying and designing innovative solutions specifically tailored to the problem of biodiversity monitoring. In the final stage of project execution we will also produce a working prototype of the system, which will be deployed and validated on selected areas of interest and made available to partner institutions of wide nature protection networks for further testing and development.