
Locally developed robot can kill mosquitoes in drains (translation)
Lianhe Zaobao, 28 Dec 2016
Researchers at the Singapore University of Technology and Design have developed a small robot that can enter narrow trenches to detect stagnant water and spray mosquito repellent agent, which would help with Singapore’s fight against diseases such as dengue and the Zika virus. The “VeSuRo” vector surveillance robot was developed by two SUTD researchers from the Engineering Product Development pillar, who have co-led this successful three-year research.
Current approaches to mosquito eradication mainly relies on traps and manual drain checks; the process is largely labour intensive and time consuming, and puts the environmental officer at risk of being bitten by mosquito.
Measuring 35cm long, 30cm wide, 32cm high and weighing 3.5kg, the VeSuRo is installed with a camera and an analysis programme which can not only replace manual drain inspection and detection of mosquito population, but also aims to have the automated capability of estimating amount of insecticide to spray based on mosquito population size.
Dr Mohan Rajesh Elara, who was in charge of developing the robot, said the third generation VeSuRo which will have the insecticide spraying capability, will be completed in the next three months. However, the user would still need to use a mobile phone application to control the operation, tracking and spraying functions. The automation of insecticide spraying capability would be in the future. Besides having conducted trials in the SUTD campus and at Pulau Ubin, he revealed that SUTD is also working with the National Environment Agency (NEA) to leverage on their vast data on mosquitoes in order to expedite the machine’s mosquito image recognition “learning capability”.
Assistant Prof Rajesh Chandramohanadas said: “Different species of mosquitoes have different shapes, structure and even wing span. Our vision information and pattern recognition algorithms are able to identify and classify the mosquitoes. It is able to differentiate whether it is a mosquito or some other insect as it flies across the camera lens, with an accuracy rate of 95 per cent.”
Besides Singapore’s concern with the spread of mosquito-related diseases like dengue and Zika, he pointed out that the algorithm can also identify the Anopheles mosquito that transmits Malaria and the Culex mosquito that spreads the West Nile virus. Next, they hope to be able to identify male mosquitoes, including the female mosquitoes that bite humans.
The VeSuRo can be mounted with different wheels that allow for five locomotion modes, be it spinning on the spot, wading or going up or down slopes, and can be easily configured into any of the five forms to access different drain conditions. The VeSuRo robot was 3D printed at SUTD, using ABS plastic that is a strong and hard material.
Mohan said: “What’s unique about our research is that each stage of the product can be brought to market, but new functions will be added to each new stage of research, combined with the new technology from different robots. For example, we are able to generate online mosquito density maps through real-time computer vision based on automated mosquito detection and classification strategies.”