Abstract: Technological Innovation in Mosquito Monitoring (Species, Genus and Age)
Process automation, generate surveillance data in real time, reduction of logistics costs, human resources, time optimization and correct prioritization of hot spots.
IoT sensor application for mosquito monitoring: Remote and automatic classification of mosquitoes: counts, genus, species, sex and age. Provides near real time information. Savings of 80% in field operations costs Embedded in existing applications based on satellite and citizen science data modelling, already being used by health agencies: VECMAP, Mosquito Alert
Lab results so far- Data analysis using machine learning techniques allow us to separate the different variables tested (species, sex and age) with a range accuracy from 61,3% to 99% .
Continuous studies with different mosquito species Species list (target species) Aedes aegypti Aedes albopictus Aedes caspius Aedes vexans Anopheles atroparvus Culex laticinctus Culex pipiens Culex theileri Culex hortensis Culiseta longiareolata Chironomidae
Variables list Species Sex Age (2-14 days) Temperature (18-28ºC) Size (large, small) Parity (nulliparous, parous, blood fed, gravid) Nutritional status (starved, sugar fed, blood fed)
Value for money: Our sensor provides real-time data, or a daily report in the worst case scenario. To provide daily data using SoA traps, users would have to increase their field operation costs 15-20 times.
Scalability and Flexibility: The sensor is designed to be modular in field deployment, and is highly robust as proven with field trials running since 2018 ( at least 5 year life time). It can de easily adapted to different types of traps: adult traps, gravid traps and ovitraps.
Power needs: Sensor consumes less power than the fan of commercial adult traps. In the case of adult traps, the sensor can be powered the same way as the trap itself: power grid or batteries.
Communications: Mobile networks (2G, 3G, 4G, 5G…) with world coverage IoT SIM. Other technologies can be used like WiFi, BlueTooth and LPWAN .
Easy to use: The sensor is easily attached to commercial traps. As soon as it is powered, it will automatically initiate itself, run diagnostics, and start collecting and transmitting data. The sensor does not even a turn on-off button.
IoT standards: OGC-SWE- (Open Geospatial Consortium - Sensor Web Enablement), IEE1451- (Smart Transducer Interface for Sensors and Actuators).
Mosquito classification: The sensor is trained in lab trials with known species of mosquitoes. This enables creating Machine Learning algorithms for those species. Nowadays the sensor is trained for Culex pipiens, Aedes aegypti and Aedes albopictus.
Accuracy: With the algorithms, we make the identification of only flying insects (we reject any other object that is not a flying insect, like suspended dust particles, pollen, rain, leaves, etc.). The sensor can be used to detect Genus, Species and Sex in areas colonized with mosquitoes for which the sensor has been trained. If the sensor is placed in an area colonized by “unknown” species, the sensor is still capable of classifying Genus and Sex.
