Design and Development of Energy Efficient Algorithms for Smart Agriculture

K V V, Durga Prasad and P, Rajalakshmi (2018) Design and Development of Energy Efficient Algorithms for Smart Agriculture. Masters thesis, Indian Institute of Technology Hyderabad.

Text
Thesis_Mtech_EE_4154.pdf - Submitted Version
Restricted to Repository staff only until 2023.
Download (14MB) | Request a copy

Abstract

This thesis addresses some of the problems involved in the IoT for agriculture. Agriculture being the backbone of Indian economy and India being largest producer of many daily needs in world. The invovlement of IoT in agriculture is not very prominant in India. There are two reasons because of which the IoT Applications of agriculture are lagging behind in India these are: 1.the lack of techological awareness in the farmers and 2. the cost of the IoT devices. Most of the Indian farmers are poor and cost plays crucial in the selection of technolgy for agriculture for them. Taking this cost factor and low power requirement of the IoT devices we have proposed a low cost IoT network for smart agriculture using Zigbee. The IoT deiveces consists of different sensors to monitor various parameters. One of the problem with IoT network deployed �eld is the internet connection, in rural area cellular network is the only source of internet which is also very poor inorder to address this we have proposed a LoRa based network as backbone to the sensor network for agriculture. The LoRa is a long range radio which has a communication range of 2km-3km distance in line of sight. The proposed network has been analysed in terms of cost and power. For a large agriculture �eld we need a larger size network which involve more number of nodes. The data gathering in such netwoks needs to be efficient for longer network lifetime. For large network we have proposed a clustering and routing algorithm which have been implemented in hardware we analysed the energy consumption of each node for the implemented algorithm in the network formation. One of the important need of IoT for agriculture is the sensors. Even though ther are many sensors available most of them either very expensive or not suitable for online monitoring. Soil moisture plays a critical role in agriculture it gives many beni�ts to the farmers in terms of saving water to proper and automated irrigation to �eld. Proper irrigation will lead to good yield to the farmers. Eventhough there are ben�ts of soil moisture most of the commercially avilable sensors which are accurate are expensive and not a�ordable to farmers. To addres this we have designed and fabricated a low cost capcitance based soil moisture sensor with a simple signal conditioning circuit. This sensor has been claberated using gravimetric method. The designed sensor is also tested in the real �eld deployment and the estimated values have been veri�ed with the gravimetric method values. The proposed sensor has a linear response with the soil moisture and good sensitivity. The sesor also has good penetration depth which gives us larger sensing volume. The issue with the proposed sensor is that the sensor needs to be calibrated for every soil type where it is to be uesd. In order address this issue we have suggested an idea in which model the sensor in lumped circuit models. The agriculture is a an IoT enviroment which involves many sensing parameters. Sensing many parameters involves multiple sensors. In such cases most of the energy is consumed in transporting this large data. So transmitting larger packet will put more burden on the battery powered nodes which in turn effects the overall network lifetime. In order to reduce the burden many compressive sensing based data aggregation methods have been proposed. In last chapter we have tried to compare the various Compressive sensing based data aggregation methods in terms of the latency in data gathering procedure. This analysis helps us in selecting the right method for the desired data capacity, latency and complexity.

[error in script]

IITH Creators: