Search

Fishermen are completely isolated from the sea shore while they are fishing in the ocean. They travel more than 120 km into the ocean for fishing. Mobile phone coverage exists only till 15 km from the shore. In such scenario, passing information like bad weather alerts to them is nearly impossible. Due to which hundreds of fisherman could lose their lives in the ocean. ‘OceanNet’ is an affordable solution developed to solve this problem in communication. The system comprises ad-hoc infrastructure networks that uses long-range Wi-Fi solution and enables offshore connectivity for fishing boats. The long range CPE deployed in the boats used directional antennas. As the boat moves, the orientation of this directional antenna changes leading to loss of connectivity. In this paper, we propose a solution to solve this problem by developing an automated antenna reorientation system that maintains the alignment between the CPE and the base station and ensures good connectivity. We also discuss on the outcome of this developed solution that solve the problem with the antenna alignment.

Real-time weather alerts will be very helpful to marine fishermen to avert the dangers from storms. Problems happen when the alerts are issued when the fishermen are already in sea. OceanNet, an affordable communication infrastructure for marine fishermen developed by our research center, will be used for this purpose. Weather stations installed on fishing vessels will perform fine-grained monitoring of the sea weather and alert the fishermen in real-time. The weather station will consist of low-cost sensors to monitor temperature, light, salinity, wind, humidity, etc. Towards this, we plan to develop a low-cost sensor for measuring the temperature of sea surface for better understanding of weather changes. Several techniques are existing to measure the sea surface temperature but most of them are expensive with less accuracy. This paper presents a survey of the available techniques and their comparison. A cost-effective technique based on thermopiles is selected for our prototype based on this survey.

In the recent years, natural disasters have become frequent due to many causes such as global warming. Flooding is one of the most common natural disasters. In this work, we introduce a flood monitoring system based on Computer Vision. The system determines the depth of the water from images captured using smartphones. The average human height is used as reference to estimate the water level. The human face is classified based on gender and age group so that the average human height of the corresponding category can be used in estimation of water depth. The data set for the system is preprocessed to mitigate the effect of poor lighting conditions, occlusions and alignment. The ground-truth validation is done using images with known water depth to determine the accuracy of the system developed.

Our research center has developed and deployed a low-cost communication infrastructure, OceanNet, for marine fishermen using long range Wi-Fi for backhaul. In the current deployments, the transceiver on the fishing vessel is not enclosed and is therefore exposed to the atmospheric salinity. Even though the equipment is built to last under these harsh conditions, it is a worthwhile exercise to assess the impact of keeping it inside an enclosure. This work analyses the effect of placing the transceiver inside one cm thick PVC enclosure in terms of the additional path loss. The variation in path loss is measured using an experimental setup and is also studied using a path loss model. The results show that an additional 2 dB path loss is incurred due to the enclosure. Therefore, it is decided not to use an enclosure in our deployments for now.

Marine fishermen need weather updates in realtime. Currently, there is no affordable communication mechanism for them from the sea. OceanNet is a solution developed by our research center to serve this purpose. It is proposed to monitor oceanic microclimate using OceanNet by installing weather stations and sensors on boats and to issue realtime weather alerts to fishermen. A low-cost solar irradiance sensor is one of the sensors to be built in-house to monitor the weather changes. Accurate measurement of solar irradiance will aid research in the area of climate change adaptation. This paper describes the survey of solar irradiance sensors and measurement techniques conducted for this purpose. The review covers three types of commercial instruments and five different research prototypes for solar irradiance measurement. A suitable technique is formulated for our prototype based on this survey.

Due to its significance in areas such as traffic surveillance, autonomous navigation, transportation management and law enforcement, vehicle classification has become a popular field of study. We plan to use vehicle classification for estimation of flood depth using images with partially submerged vehicles. The aim of this paper is to provide researchers with an overview of different machine learning algorithms used for vehicle classification using Convolutional Neural Networks (CNN) architecture. Different CNN algorithms have been used for classification of vehicles efficiently. This work reviews and compares the various classification algorithms for detecting the type as well as the model of a vehicle.

Our research center has successfully developed and deployed OceanNet, a novel high-speed communication infrastructure for marine fishermen which has a range of 60-70 km from the shore. We plan to use this network for coastal oceanography and weather monitoring by installing an IoT of low-cost weather stations equipped with multiple sensors on fishing vessels. Towards this, we are developing a suite of low-cost and yet feature rich sensors. Ultrasonic anemometer is one such sensor. A survey of the commercially available anemometers revealed that they are too expensive for our use and also not fully featured to meet our requirements. Various research prototypes were also surveyed to get a realistic assessment of the price point and the feature set that we can target for our prototype. This work presents a detailed account of the survey of the available anemometers with a comparison of their key parameters.