A User-Centered Security Approach to Create an IoT Based Multi-Layered Fog-Cloud Architecture for Data Optimization in Raised Bed Farming

A User-Centered Security Approach to Create an IoT Based Multi-Layered Fog-Cloud Architecture for Data Optimization in Raised Bed Farming

Rishi Raghu Vijayashree JayaramaanJayashree Jayaraman Sai Surya Varsith Nukala Carlos Enrique Montenegro-Marin 

School of Computer Science and Engineering, Vellore Institute of Technology, Vellore 632014, India

Facultad de ingeniería, Universidad Distrital Francisco José de Caldas, Bogota 110231, Colombia

Corresponding Author Email: 
30 August 2022
16 October 2022
31 December 2022
| Citation



Cloud computing and Internet of Things (IoT) are popularly intertwined to create a network of smart systems, especially in the field of Agriculture. As agricultural industries rise, augmented use of heterogeneous communication and sensing on the field would be needed. For all this data to be transferred and processed on the cloud efficiently, low latency rates, consistently high bandwidths, minimal congestion, etc. would be needed, which, is still a complication with the current cloud computing models. This paper proposes a multi-layered fog architecture that detects outliers in the data received from the sensing environment based on three categories: Classification, Isolation, and Clustering and then aggregates it before sending it to the cloud. The architecture works closely in a user-centered design approach that connects the farmers and analysts to the fog allowing them to create an automated agricultural system. With the help of Fog, processing abilities are brought closer to the data source which reduces the load on cloud resources, thereby making the overall system a lot more efficient and secured. This paper also presents a prototype of the interface that can be used to monitor and control IoT devices on the field as well as define fuzzy rules for the agricultural system.


raised bed farming, IoT, fog computing, cloud computing, outlier recognition

1. Introduction

IoT is an interconnection of various devices that use embedded technology to sense, communicate, and interact with their internal and external environments [1]. IoT has emerged as a megatrend for next-generation technology that has the potential to impact the entire business sector by providing enhanced connection of end devices, systems, and applications [2].

As a result of the emergence of Industrial IoT, far more advanced sensors are being used in agriculture. A cellular/satellite network is currently used to connect the sensors to the cloud. The real-time data from the sensors allow us to make better decisions [3]. The advancement of IoT technology in agriculture operations has enabled the use of sensors at every stage of the agricultural process, such as how much time and resources a seed requires to mature into a completely grown crop. Every object that can be operated over the internet is referred to as a device [4]. Wearable IoWT (Internet of Wearable Things) gadgets like smart watches and home management solutions like Google Home have made IoT devices very popular in consumer markets [5]. By 2020, it is expected that over 30 billion gadgets will be connected to the Internet of Things. The uses of IoT in agriculture are aimed at traditional farming [6].

Applicability of IoT in Agriculture:

Smart farming is a high-tech and efficient approach to agricultural and food production that is both sustainable and efficient. It's a way of incorporating connected gadgets and cutting-edge technology into agriculture. Smart farming relies largely on the Internet of Things, which reduces the need for farmers and growers to perform physical labor and so boosts productivity in every way possible [7-9]. IoT-based Smart Farming improves the overall Agriculture system by monitoring the field in real-time [10, 11]. Thanks to sensors and interconnectivity, the Internet of Things in Agriculture has not only saved farmer’s time but has also reduced resource waste such as water and power. It keeps track of a range of characteristics, including humidity, temperature, and soil, and delivers feedback [12].

Figure 1. Motive of raised bed farming

Robots, drones, remote sensors, and computer imagery, along with machine learning and analytical tools, are also used in farming to monitor crops, deliver data to farmers for sensible farm management plans, saving time and money. Cloud computing has a lot of potential in the agriculture business. By incorporating IT into the traditional farming system, it can be improved. Motive of raised bed farming is shown in Figure 1.

Mocau et al. [8] has proposed a cloud-based infrastructure to supplement traditional farming management systems, which uses cloud-based architecture to monitor farm management. These days cloud computing is an extensively used service across the world, but there are major problems with the clouds. Here we are going to discuss some important problems.

1.1 Challenges and limitations of cloud

These days we can see traffic on cloud environments is constantly increasing becoming a big problem which in turn may lead to network congestion and many other problems [13]. Since Cloud data is accessible through public internet platforms, there are many chances that data is intruders may get easy access to valuable and confidential data of the customers. Even though numerous efforts are being done to secure cloud computing systems from many vulnerabilities and data breaches, but still, it is being a challenging task for IT security professionals. Cloud Service latency is becoming one of the alarming problems [14]. Latency depends upon the geographical location of the end-users where users far away from cloud servers may experience poor services like delay in transmission of data. Without the internet, we can’t even imagine using cloud services. Hence, cloud services aren’t accessible at any internet connectivity locations. And they are highly time-sensitive at poor no internet connectivity areas.

1.2 Transformation towards IoT and fog computing

At today’s pace of technological revolution related to the Internet of Things where many objects need to be connected to the internet and so much data needs to be analyzed and processed [15]. Transmission of all these data to the cloud would be difficult because of the increase in latencies and traffic within the network; this also requires high processing capacity and bandwidth [16]. So, the Fog Computing paradigm shown in Figure 2 is introduced to reduce the above problems by supporting the cloud computing architecture.

Fog computing aims to work on a decentralized approach that helps to reduce data traffic problems in centralized cloud computing [17]. Fog layer is represented by a sensor node and actuators where it uses algorithms to filter sensor data, tasks like cluster analysis, alert management, etc., are performed to reduce the computational load on the cloud [18]. Fog computing can solve many IoT challenges like resource-constraint, latency constraints, IoT security challenges, Network bandwidth constraints, [19]. Fog computing reduces the latencies by performing all computational tasks close to the end-users [20].

1.3 The interface of fog with the cloud and IoT

The fog layer serves as a bridge between the IoT and the cloud. It is responsible to communicate among the two layers and acts as the common interface layer between them as shown in Figure 3.

Figure 2. Fog computing architecture

Figure 3. Interface of fog with cloud and IoT

Depending on the type of communication, various types of interfaces are present between each layer such as Fog-Fog Interface, Fog-IoT Interface, Fog-Cloud Interface, Cloud-IoT Interface [21, 22]. The fog layer should be used to hold real-time data that requires quick processing. Similarly, mass storage and computational data that requires a lot of processing power should be stored in the cloud [23]. The difference between cloud and fog computing is given in Table 1.

1.4 Seven-layer fog computing architecture

The seven-layer fog computing architecture contains seven layers as shown in Figure 4. The components are organized into many categories on the basis of their usefulness, that is determined by the layer [24, 25]. These features will allow sensor nodes to interconnect with fog and the cloud [26].

Table 1. Comparison of cloud and fog computing


Cloud computing

Fog computing




Real time interactions






Distance between client node and server node

Multiple hop

One hop

Proximity and geographic coverage



Target users

Internet users

Mobile and resource constrained users




Resource management



Location Awareness



Location of Service

Within internet

At the edge of local network

Security measures


Hard to define




Distance to end devices



Connectivity and communication

IP-based only

IP and non-IP



Servers, switches, gateways, access points, Routers

Bandwidth required



Content generator


Sensors and devices


Data center

single server to a micro-data center


3 models (PaaS, IaaS,SaaS)



High probability

Very low probability

Deployment cost



Figure 4. Layered architecture of fog computing

Transport Layer

All the data which is pre-processed at the fog level, needs to be transferred to the cloud. The transport layer is responsible for updating all the pre-processed data to the cloud while maintaining security and latency [27]. Due to the efficient pre-processing done at the previous layers, the load on the transport layer is low and as a result quite fast and secure. This layer is additionally responsible for all the computational requirements requested by various applications at the fog level [28].

Security Layer

All security issues, such as encryption and decryption of data that is exchanged inside the layers, are addressed as indicated by the name. To preserve security, all communications between IoT devices and the cloud must be encrypted; the encryption component encrypts all connections from and to IoT devices and the cloud. Fog computing, like cloud computing, was supposed to be more of a utility computing notion [29].

Resource Management Layer

This layer is in charge of allocating resources, scheduling them, and dealing with energy-saving disputes. The layer itself comprises various components out of which one is the reliability component which preserves the consistency of scheduling an application and its resource allocation [30].

Temporary Storage Layer

The temporary storage layer is responsible for storing all the pre-processed data through virtualization [31]. It comprises an information backup component that guarantees the availability of knowledge and reduces the loss of information [13].

Pre-Processing Layer

This layer consists of sub-layers, which works on data analysis pertaining to the respective fog layer. Firstly, all the gathered data is analyzed and filtered, data trimming and reconstruction is done.

Monitoring Layer

The monitoring layer is the one that keeps track of the system's performance and reactions at all times. We can choose the suitable resources during the operation thanks to system monitoring. Most applications run on smart transportation systems which creates a high chance of facing resource unavailability for efficient storage or computation on a Fog device. To tackle situations like these, the Fog device and the servers will reach out to their neighboring devices and components [32].

Physical Layer

The physical layer is made up of all the virtual and physical sensors present in our architecture, where any and every data generation device could comprise any of the groups. This layer is answerable for all the sensing of relevant data before it's forwarded to the other layers.

1.5 Communication protocols

Based on our proposed architecture we require a multi-level communication system that connects Fog computing with cloud and IoT [33]. To connect the various layers of IoT, a lot of different communication protocols are available such as Internet Protocol Version 6 (IPv6), Z-Wave and Near Field Communication (NFC), Low power Area Networks (6LoWPAN), Bluetooth Low Energy (BLE), ZigBee, etc. [34].

The most widely used IoT protocol to connect networks on Fog is ZigBee (IEEE 802.15.4 standard) mainly due to its long battery life and low duty cycle. Networks that run on this protocol use a mesh topological system and mainly operate at 2.2 GHz -900 Mhz unlicensed frequency bands. Sensors deployed within a range of 100-150m are used for farm monitoring and management using Zigbee technology [9]. With the help of protocols like ZigBee data is efficiently gathered from these sensors and transferred to a cloud data center for further analysis [35, 36].

In smart farming, for long-range communications GSM cellular network is also used especially for smart irrigation and environmental monitoring. For the communication between gateways, the MQTT (Message Queuing Telemetry Transport) protocol is considered as the reference standard for IoT communication [37]. MQTT is a messaging protocol suited for working with limited power computing and connectivity.

2. Literature Review

Table 2 summarizes various research works on IoT, methods, devices and models used.

Table 2. Review of research contributions in the area of IoT

Research Domain


Research Contribution

Method/ Device/ Models used

Environment Monitoring


IoT based soil monitoring for farming

Wireless sensors


IoT for air pollution monitoring system

Gas sensor and IoT


smart environment monitoring

Heterogeneous sensors


Air quality monitoring

Geomatics sensors and IoT


Aqua farming and energy conservation

pH, and temperature sensor


Climate monitoring

LoRa technology and sensor network


IoT protocols for acoustic monitoring in the marine environment

WSN and IoT

Agriculture Monitoring


Smart Agriculture

Pest detection, crop status, irrigation, Soil preparation


Smart Agriculature

Deep reinforcement learning, machine learning


Precision Agriculture

Cloud computing, IoT, edge computing



AI, block chain and edge computing


Agriculture data analytics

The combination of Data analytics and IoT is enabling smart agriculture.


Smart irrigation

Wireless sensor network, sensor node, and irrigation


Pest control

For pest monitoring, IoT and deep learning with global and local features are used.


Leaf area index

synthetic aperture radar (SAR) images

Air Quality Monitoring


Detecting Water contamination

IoT and machine learning


Monitoring Water quality

IoT with smart sensors


Investigation of Water contamination

Big data and SVM


Lagoon water

Machine learning and image analysis


Water quality

Neural network


Indoor air quality



Intelligent air quality system

UV light, AI and sensors


Air quality characterization

Heterogeneous sensors, AI


Air pollution

Gas sensors


CO2 monitoring

IoT and cloud technologies

3. Proposed Work

The proposed architecture contains four layers namely: Things layer, fog layer1, fog layer 2 and cloud platform as shown in Figure 5.

Things Layer

The Things layer includes devices such as temperature sensors, Accelerometer sensors, Asset Monitoring, embedded systems for light, humidity, and moisture. Things are a virtual representation of all of the resources that can be delivered to different subsystems. The devices work on the field to capture and record the necessary data required for automating the system and communicating with Fog and cloud layers. Data captured from this layer is extremely essential for filtering and Outliner recognition that takes place in fog layer 1. At the Things layer, the fundamental communication protocol is ZigBee (IEEE 802.15.4 standard). Due to its powerful battery life and low-cost cycle, it helps us to maintain an efficient latency between the fog layers, and all the data captured by the sensors and actuators are flexibly transferred to the cloud center.

Fog layer1: Outlier Recognition

The data from sensors of our rides bed farming like water sensors, soil sensors, pH sensors of water and soil, electrical conductivity sensors of soil and water, and also the data of actuators are collected and send to the first layer of fog computing [49]. There are chances that these data may contain outliers. These outliers are caused due to some battery damage problems of sensors and false readings [50]. Outliers in the data may reduce the accuracy and may lead to wrong decision-making of the system. So, in the fog layer, we use some outliner recognition techniques which use machine learning algorithms to remove outliers in the data [51]. This helps fog layer 2 in proper status generation and event detection. And finally, that helps cloud computing systems to predict future crop yield and growth. Here we use some well-known methods for outlier detection they are classification-based, isolation-based, and clustering-based.

In the classification-based method, training of a classification model using a dataset takes place. Here we used the Support Vector Machine algorithm. This algorithm determines any instance that is falling apart from the given boundary as an outlier [52]. The isolation-based method concentrates on the separation of outliners from the rest of the data points and won’t profile the normal ones. Here Isolation Forest algorithm is used. This algorithm can find the abnormalities in the data. Clustering-based method groups data which is similar to clusters with almost the equivalent behavior [53]. Instances which don’t belong to the cluster are determined as outliers [54]. Here we use the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. This algorithm groups the instances into high and low-density clusters utilizing simplified minimum density level estimation which is based on parameters like distance radius, thresholds, etc. [55].

Fog Layer 2: Resource Management

After detecting and excluding outliers in the data at fog layer 1, that filtered data is sent to fog layer 2. Here in the fog layer 2 temporary data storage, some machine learning processes and communication with the cloud system are done. The data taken from fog layer 1 is temporarily stored in this layer. Then that data analyzed using decision tree algorithms in machine learning. Those results we got from machine learning algorithms help fog layer 2 in event generation and status detection like crop watering alerts about the farm bed regarding water content in the soil, regarding nutrition and pH levels of the soil [56].

Cloud Platform

The data that crossed fog computing layers 1 and 2 reaches cloud services. Here in the cloud platform that data from sensors and actuators gets analyzed using neural network algorithms. And finally, users can get predictions about the crop yield, also suggests users the techniques to improve crop yield and identification of the crop diseases takes place. crop growth and other parameters will be visualized using data visualization techniques and will be shown to the users.

The proposed architecture used 3 different protocols at each layer of the architecture. Zigbee protocol is used for communication between sensor nodes and actuators [57]. LoRa protocol is used in between Fog layer1 and Fog layer2 to send data after outlier recognition [58-66]. And finally for resource management in between fog layer2 and cloud platform we used the MQTT protocol. The process flow of our proposed architecture is shown in Figure 6.

Figure 5. Proposed architecture

Figure 6. Process flow of proposed architecture

The process flow of our proposed architecture goes on in about 5 levels. These stages work with each other to systematize data between various fog and cloud layers to communicate with the end-user efficiently and seamlessly. The 5 levels are:

IoT sensors: A bunch of sensors of various types is used in the initial stage to sense and record data. Mainly 2 types of sensors are used in the proposed architecture first of which is Water-type sensors are responsible for calculating water pH, electrical conductivity, water temperature. The second is Soil Type sensors which take part in calculating the pH and the temperature of the soil, and its conductivity.

Outlier Recognition: This stage is essential for failure identification and outliers’ isolation from correct data. There can be a lot of reasons due to which erroneous results might be generated while transferring data from the sensors to the cloud such as low battery power, loose wiring, etc. It consists of 3 categories: Clustering, Isolation, and Classification. The clustering-based approach uses DBSCAN which segregates the entire dataset into high and low-density clusters [52]. The isolation-based approach uses Isolation Forest (IF) which is capable of identifying anomalies from a dataset it works on the principle of recursive splitting of attribute value trees to isolate data points [51]. The classification-based approach uses a Support Vector Machine (SVM) which helps to identify if any instances falling out of the boundary is a part of an outlier.

Data aggregation: This stage deals with summarizing all the data filtered out from the outlier recognition stage into categories and sections, this is to prepare the data for cloud transfer. The data is also used for Status generation and event detection from which then the data is transferred to the cloud.

Status generation and Event detection: This phase works on the fog node to carry out machine learning processes on different parameters and lets the farmer decide smart irrigation rules to optimize farming. This also returns the status of all the crops in the field. For instance, if the sensors in the field detect low water content, they generate a status regarding the low water in the field and automate the sprinklers to water the crops. Earlier, smart irrigation was controlled by time schedules. Now, with the help of this, it considers data sensors to see if decision trees can be used to optimise watering and growth control. Data is then updated and sent again for data aggregation from where it is sent to the cloud.

Cloud/End-users: With the use of neural networks present at the Cloud layer, all the data from the aggregation layer is sent and processed here for future storage. The result is that only the essential and processed data is sent to the cloud which reduces the latency thereby increasing the speed of data transfer and data processing.

4. Prototype: User-Centered Design Approach

This paper also proposes a new application prototype to provide essential services for the best management of agricultural fields. This prototype involves farmers and analyst as shown in Figure 7.

The motive behind our architecture was to create a system that delivers efficient results and analysis at the same time, reduces work for the users operating it. User-centered design or UCD is a way of defining a design process where the stakeholders (Farmers, Agriculturalists, Analysts) control how the system automates with the fog and cloud layers to deliver data. The process was designed with keeping the users at the core which is why it is essential to create a user-friendly process that the farmers, agricultural specialists, and analysts can trust and understand. The user-centered design approach works based on subjective speculations about the user, their personality, and their behaviour. Sensors, actuators, and the other IoT devices at the things layer perform tasks that automate to optimize production in the field. These devices integrated with other smart devices work to enhance interaction between the user and the fog layers.

Figure 7. Communication between farmer and analyst

The design approach consists of 2 parts: The first is in respect to the farmers and the agricultural specialists who work in close relation with the things layer. Here the user gets control over how the fuzzy system rules should be defined for the raised bed fields and gets complete control over the sensors, actuators, and the automated equipment. This is done through a smart device present on the farmer’s side. The second is based on feedback and results returned after data aggregation. At any point of the day, with the help of various data visualization techniques, the user can get a complete analysis of all the results generated from the fog layers. And based on the results, the rules and settings for the sensors can be adjusted accordingly. Since data from the cloud would also be represented here, data analysts can also take part in controlling or suggesting fuzzy rules for the system. With this approach, we can put the intended users of a system at the center of its design and development. The stages of this process are carried out in an iterative cycle, meaning the entire process is repeated until the project’s usability objectives have been achieved.

Figure 8. Interface of the dashboard and the cloud data analysis

Figure 8 shows the interface of the dashboard and the cloud data analysis. For the ease of farming and monitoring, an interface that gives out daily analysis and visualization is essential. In the dashboard, critical information such as soil moisture, temperature, humidity is displayed and updated daily. Based on the sensors and actuators configured on the field data is aggregated and the summary is displayed on the screen. Apart from this, the user can also view statistics and visualization of all the crops and their progress over time. As an example, we added a visualization graph for the crop yields and soil quality over one week. With this, the farmer can keep a check on the progress of the crop and can change the necessary fuzzy rules to get better results.

The user stays in control of the entire system throughout the architectural process. The farmer based on data depicted on the dashboard and all the data received can control the necessary sensors and actuators in the field as in Figure 9. The interface allows several various options such as smart irrigation where the user can choose the speed, pressure, and interval of water that the sprinkler shoots. Options like timer and target moisture are also available by which the system automates the need of the user. For example, if the user sets the target Moisture level as 4mm for the entire field, the sprinkler will shoot water till the sensors in the ground detect a moisture level of 4mm or above. Through the use of smart devices, the farmer along with the guidance of agriculturalists can define fuzzy rules for their system. Rules are a form of automated instructions given by the user on how the system should operate.

Figure 9. Interface for smart irrigation

If the user needs to automate fertilizer sprays which only activate when met some specific conditions. When the users need to, they can simply add a new rule to the system and disable it whenever they want. The entire prototype of the interface was designed by keeping the users at the center such that it becomes easy for them to grasp and get comfortable with the automated system.

5. Results and Analysis

Simulation is used to evaluate the usefulness of the proposed architecture. Simulations have been performed in a network composed by fifteen sensor nodes, one gateway node and the Cloud platform. The amount of data stored to the cloud in bytes is evaluated in three different scenarios: Cloud Computing architecture, 1-tier Fog architecture and 2-tier Fog architecture. The amount of data stored to the cloud in one month is shown in Figure 10.

Figure 10. Comparison of the amount of data stored in different architectures

The figure illustrates the decrease in the amount of data stored in the Cloud due to outlier recognition and filtering in fog layer. Accuracy of the three techniques: SVM, DBSCAN and IF, used for outlier recognition is given in Table 3.

Table 3. Comparison of accuracy of various outlier recognition techniques

Outlier recognition techniques

Accuracy per-day dataset

Accuracy per-month dataset










Figure 11. Comparison of accuracy of various outlier recognition techniques

Accuracy of the three techniques used for outlier recognition is evaluated using per-day and per-month datasets, and is shown in Figure 11.

It is evident from the figure that DBSCAN technique recognizes the outliers more accurately when compared to SVM and IF.

6. Conclusions

In agricultural countries, a smart agricultural field monitoring system might be quite useful. This approach can aid in the accurate management of agricultural land. This technique keeps water from being wasted. As part of this paper's future study, more sensors with more data analysis could be developed. The proposed architecture compares three outlier recognition techniques and from the analysis it is evident that DBSCAN technique recognizes the outliers more accurately when compared to SVM and IF. As the data from sensors are subjected to outlier recognition and filtering, the total amount of data sent to the cloud is reduced. This in turn balances the computational load by distributing the business logic between several layers. As a result, significant cloud resources are saved. The proposed architecture also reduces waiting time by processing the real time data locally near the field without sending the data to the cloud and also uses actuators which sends notification based on the analysis.


[1] Khan, R., Khan, S.U., Zaheer, R., Khan, S. (2012). Future internet: the internet of things architecture, possible applications and key challenges. In 2012 10th international conference on frontiers of information technology, Islamabad, Pakistan, pp. 257-260. https://doi.org/10.1109/FIT.2012.53

[2] Chen, Y., Lee, G.M., Shu, L., Crespi, N. (2016). Industrial internet of things-based collaborative sensing intelligence: Framework and research challenges. Sensors, 16(2): 215. https://doi.org/10.3390/s16020215

[3] Atzori, L., Iera, A., Morabito, G. (2010). The internet of things: A survey. Computer Networks, 54(15): 2787-2805. https://doi.org/10.1016/j.comnet.2010.05.010

[4] Junior, F.M.R., Kamienski, C.A. (2021). A Survey on Trustworthiness for the Internet of Things. IEEE Access, 9: 42493-42514. https://doi.org/10.1109/ACCESS.2021.3066457

[5] 1Atlam, H.F., Alenezi, A., Alharthi, A., Walters, R.J., Wills, G.B. (2017). Integration of cloud computing with internet of things: challenges and open issues. In 2017 IEEE International Conference on Internet of Things (iThings) and IEEE Green Computing and Communications (GreenCom) and IEEE Cyber, Physical and Social Computing (CPSCom) and IEEE Smart Data (SmartData), pp. 670-675.

[6] Zyrianoff, I., Heideker, A., Silva, D., Kleinschmidt, J., Soininen, J.P., Salmon Cinotti, T., Kamienski, C. (2020). Architecting and deploying IoT smart applications: A performance–oriented approach. Sensors, 20(1): 84. https://doi.org/10.3390/s200100846

[7] Lakhiar, I.A., Gao, J., Syed, T.N., Chandio, F.A., Buttar, N.A. (2018). Modern plant cultivation technologies in agriculture under controlled environment: A review on aeroponics. Journal of Plant Interactions, 13(1): 338-352. https://doi.org/10.1080/17429145.2018.1472308

[8] Stafford, J.V. (2000). Implementing precision agriculture in the 21st century. Journal of Agricultural Engineering Research, 76(3): 267-275. https://doi.org/10.1006/jaer.2000.0577

[9] Augustin, A., Yi, J., Clausen, T., Townsley, W.M. (2016). A study of LoRa: Long range & low power networks for the internet of things. Sensors, 16(9): 1466. https://doi.org/10.3390/s16091466

[10] Yan-e, D. (2011). Design of intelligent agriculture management information system based on IoT. In 2011 Fourth International Conference on Intelligent Computation Technology and Automation, pp. 1045-1049. https://doi.org/10.1109/ICICTA.2011.262

[11] Jayaraman, P.P., Yavari, A., Georgakopoulos, D., Morshed, A., Zaslavsky, A. (2016). Internet of things platform for smart farming: Experiences and lessons learnt. Sensors, 16(11): 1884. https://doi.org/10.3390/s16111884

[12] Riquelme, J.L., Soto, F., Suardíaz, J., Sánchez, P., Iborra, A., Vera, J.A. (2009). Wireless sensor networks for precision horticulture in Southern Spain. Computers and Electronics in Agriculture, 68(1): 25-35. https://doi.org/10.1016/j.compag.2009.04.006

[13] Hsu, T.C., Yang, H., Chung, Y.C., Hsu, C.H. (2020). A Creative IoT agriculture platform for cloud fog computing. Sustainable Computing: Informatics and Systems, 28: 100285. https://doi.org/10.1016/j.suscom.2018.10.006 

[14] Ray, P.P. (2017). Internet of things for smart agriculture: Technologies, practices and future direction. Journal of Ambient Intelligence and Smart Environments, 9(4): 395-420. https://doi.org/10.3233/AIS-170440

[15] Aazam, M., Huh, E.N. (2014). Fog computing and smart gateway based communication for cloud of things. In 2014 International Conference on Future Internet of Things and Cloud, pp. 464-470. https://doi.org/10.1109/FiCloud.2014.83

[16] Vaquero, L.M., Rodero-Merino, L. (2014). Finding your way in the fog: Towards a comprehensive definition of fog computing. ACM SIGCOMM Computer Communication Review, 44(5): 27-32. https://doi.org/10.1145/2677046.2677052

[17] Bonomi, F., Milito, R., Zhu, J., Addepalli, S. (2012). Fog computing and its role in the internet of things. In Proceedings of the First Edition of the MCC Workshop on Mobile Cloud Computing, pp. 13-16. https://doi.org/10.1145/2342509.2342513

[18] Mouradian, C., Naboulsi, D., Yangui, S., Glitho, R.H., Morrow, M.J., Polakos, P.A. (2017). A comprehensive survey on fog computing: State-of-the-art and research challenges. IEEE Communications Surveys & Tutorials, 20(1): 416-464.

[19] Aazam, M., Zeadally, S., Harras, K.A. (2018). Fog computing architecture, evaluation, and future research directions. IEEE Communications Magazine, 56(5): 46-52. https://doi.org/10.1109/MCOM.2018.1700707

[20] Kim, H.S. (2016). Fog computing and the internet of things: Extend the cloud to where the things are. International Journal of Cisco.

[21] Saharan, K.P., Kumar, A. (2015). Fog in comparison to cloud: A survey. International Journal of Computer Applications, 122(3).

[22] Yousefpour, A., Ishigaki, G., Jue, J.P. (2017). Fog computing: Towards minimizing delay in the internet of things. In 2017 IEEE International Conference on Edge Computing (EDGE), pp. 17-24.

[23] Mukherjee, M., Shu, L., Wang, D. (2018). Survey of fog computing: Fundamental, network applications, and research challenges. IEEE Communications Surveys & Tutorials, 20(3): 1826-1857. https://doi.org/10.1109/COMST.2018.2814571

[24] Hao, Z., Novak, E., Yi, S., Li, Q. (2017). Challenges and software architecture for fog computing. IEEE Internet Computing, 21(2): 44-53. https://doi.org/10.1109/MIC.2017.26

[25] Robert, P.C. (2002). Precision agriculture: a challenge for crop nutrition management. In Progress in Plant Nutrition: Plenary Lectures of the XIV International Plant Nutrition Colloquium, pp. 143-149.

[26] Ray, P.P. (2018). A survey on Internet of Things architectures. Journal of King Saud University-Computer and Information Sciences, 30(3): 291-319. https://doi.org/10.1016/j.jksuci.2016.10.003

[27] Ma, X., Liang, J., Liu, R., Ni, W., Li, Y., Li, R., Ma, W., Qi, C. (2018). A survey on data storage and information discovery in the WSANs-based edge computing systems. Sensors, 18(2): 546. https://doi.org/10.3390/s18020546

[28] Yassein, M.B., Shatnawi, M.Q. (2016). Application layer protocols for the Internet of Things: A survey. In 2016 International Conference on Engineering & MIS (ICEMIS), Agadir, Morocco, pp. 1-4. https://doi.org/10.1109/ICEMIS.2016.7745303

[29] Lai, C.F., Zeadally, S., Shen, J. and Lai, Y.X. (2015). A cloud-integrated appliance recognition approach over internet of things. Journal of Internet Technology, 16(7): 1157-1168.

[30] Dastjerdi, A.V., Gupta, H., Calheiros, R.N., Ghosh, S.K., Buyya, R. (2016). Fog computing: Principles, architectures, and applications. In Internet of Things, pp. 61-75. https://doi.org/10.1016/B978-0-12-805395-9.00004-6

[31] Malche, T., Maheshwary, P. (2017). Internet of things (IoT) based water level monitoring system for smart village. In Proceedings of International Conference on Communication and Networks, pp. 305-312.

[32] Talavera, J.M., Tobón, L.E., Gómez, J.A., Culman, M.A., Aranda, J.M., Parra, D.T., Quiroz, L.A., Hoyos, A., Garreta, L.E. (2017). Review of IoT applications in agro-industrial and environmental fields. Computers and Electronics in Agriculture, 142: 283-297. https://doi.org/10.1016/j.compag.2017.09.015

[33] Chikankar, P.B., Mehetre, D., Das, S. (2015). An automatic irrigation system using ZigBee in wireless sensor network. In 2015 International Conference on Pervasive Computing (ICPC), pp. 1-5. https://doi.org/10.1109/PERVASIVE.2015.7086997

[34] Hortelano, D., Olivares, T., Ruiz, M.C., Garrido-Hidalgo, C., López, V. (2017). From sensor networks to internet of things. Bluetooth low energy, a standard for this evolution. Sensors, 17(2): 372. https://doi.org/10.3390/s17020372

[35] Keller, T., Sandin, M., Colombi, T., Horn, R., Or, D. (2019). Historical increase in agricultural machinery weights enhanced soil stress levels and adversely affected soil functioning. Soil and Tillage Research, 194: 104293.

[36] Domínguez-Niño, J.M., Oliver-Manera, J., Girona, J., Casadesús, J. (2020). Differential irrigation scheduling by an automated algorithm of water balance tuned by capacitance-type soil moisture sensors. Agricultural Water Management, 228: 105880. https://doi.org/10.1016/j.agwat.2019.105880

[37] Luzuriaga, J.E., Perez, M., Boronat, P., Cano, J.C., Calafate, C., Manzoni, P. (2015). A comparative evaluation of AMQP and MQTT protocols over unstable and mobile networks. In 2015 12th Annual IEEE Consumer Communications and Networking Conference (CCNC), Las Vegas, NV, USA, pp. 931-936. https://doi.org/10.1109/CCNC.2015.7158101

[38] Zuquim, G., Costa, F.R.C., Tuomisto, H., Moulatlet, G.M., Figueiredo, F.O.G. (2020). The importance of soils in predicting the future of plant habitat suitability in a tropical forest. Plant and Soil, 450(1): 151-170. https://doi.org/10.1007/s11104-018-03915-9

[39] Maček, D., Magdalenić, I., Ređep, N.B. (2020). A systematic literature review on the application of multicriteria decision making methods for information security risk assessment. International Journal of Safety and Security Engineering, 10(2): 161-174. https://doi.org/10.18280/ijsse.100202

[40] Chen, J., Yang, A. (2019). Intelligent agriculture and its key technologies based on internet of things architecture. IEEE Access, 7: 77134-77141. https://doi.org/10.1109/ACCESS.2019.2921391

[41] Liu, S., Guo, L., Webb, H., Ya, X., Chang, X. (2019). Internet of Things monitoring system of modern eco-agriculture based on cloud computing. IEEE Access, 7: 37050-37058. https://doi.org/10.1109/ACCESS.2019.2903720

[42] Saad, A., Gamatié, A. (2020). Water management in agriculture: A survey on current challenges and technological solutions. IEEE Access, 8: 38082-38097. https://doi.org/10.1109/ACCESS.2020.2974977

[43] Colombi, T., Keller, T. (2019). Developing strategies to recover crop productivity after soil compaction—A plant eco-physiological perspective. Soil and Tillage Research, 191: 156-161. https://doi.org/10.1016/j.still.2019.04.008

[44] Farooq, M.S., Riaz, S., Abid, A., Abid, K., Naeem, M.A. (2019). A survey on the role of IoT in agriculture for the implementation of smart farming. IEEE Access, 7: 156237-156271. https://doi.org/10.1109/ACCESS.2019.2949703

[45] Chen, D., Wang, X., Zhang, W., Zhou, Z., Ding, C., Liao, Y., Li, X. (2020). Persistent organic fertilization reinforces soil-borne disease suppressiveness of rhizosphere bacterial community. Plant and Soil, 452: 313-328. https://doi.org/10.1007/s11104-020-04576-3

[46] Putjaika, N., Phusae, S., Chen-Im, A., Phunchongharn, P., Akkarajitsakul, K. (2016). A control system in an intelligent farming by using arduino technology. In 2016 Fifth ICT International Student Project Conference (ICT-ISPC), pp. 53-56. https://doi.org/10.1109/ICT-ISPC.2016.7519234

[47] Zhang, X., Zhang, J., Li, L., Zhang, Y., Yang, G. (2017). Monitoring citrus soil moisture and nutrients using an IoT based system. Sensors, 17(3): 447. https://doi.org/10.3390/s17030447

[48] Taneja, M., Jalodia, N., Malone, P., Byabazaire, J., Davy, A., Olariu, C. (2019). Connected cows: Utilizing fog and cloud analytics toward data-driven decisions for smart dairy farming. IEEE Internet of Things Magazine, 2(4): 32-37. https://doi.org/10.1109/IOTM.0001.1900045

[49] Elbamby, M.S., Bennis, M., Saad, W., Latva-Aho, M., Hong, C.S. (2018). Proactive edge computing in fog networks with latency and reliability guarantees. EURASIP Journal on Wireless Communications and Networking, 2018(1): 1-13. https://doi.org/10.1186/s13638-018-1218-y

[50] Çelik, M., Dadaşer-Çelik, F., Dokuz, A.Ş. (2011). Anomaly detection in temperature data using DB scan algorithm. In 2011 International Symposium on Innovations in Intelligent Systems and Applications, pp. 91-95.

[51] Wahir, N.A., Nor, M.E., Rusiman, M.S., Gopal, K. (2018). Treatment of outliers via interpolation method with neural network forecast performances. Journal of Physics: Conference Series, 995(1): 012025. https://doi.org/10.1088/1742-6596/995/1/012025

[52] Huang, C., Liu, D., Ni, J., Lu, R., Shen, X. (2018). May. Reliable and privacy-preserving selective data aggregation for fog-based IoT. In 2018 IEEE International Conference on Communications (ICC), pp. 1-6.

[53] Ding, Z., Fei, M. (2013). An anomaly detection approach based on isolation forest algorithm for streaming data using sliding window. IFAC Proceedings Volumes, 46(20): 12-17. https://doi.org/10.3182/20130902-3-CN-3020.00044

[54] Schubert, E., Sander, J., Ester, M., Kriegel, H.P., Xu, X. (2017). DBSCAN revisited, revisited: Why and how you should (still) use DBSCAN. ACM Transactions on Database Systems (TODS), 42(3): 1-21. https://doi.org/10.1145/3068335

[55] Liu, Y., Wang, H., Zhang, H., Liber, K. (2016). A comprehensive support vector machine-based classification model for soil quality assessment. Soil and Tillage Research, 155: 19-26. https://doi.org/10.1016/j.still.2015.07.006

[56] Boutaba, R., Salahuddin, M.A., Limam, N., Ayoubi, S., Shahriar, N., Estrada-Solano, F., Caicedo, O.M. (2018). A comprehensive survey on machine learning for networking: Evolution, applications and research opportunities. Journal of Internet Services and Applications, 9(1): 1-99. https://doi.org/10.1186/s13174-018-0087-2

[57] Heđi, I., Špeh, I., Šarabok, A. (2017). IoT network protocols comparison for the purpose of IoT constrained networks. In 2017 40th International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO), pp. 501-505.

[58] Al-Sarawi, S., Anbar, M., Alieyan, K., Alzubaidi, M. (2017). Internet of Things (IoT) communication protocols. In 2017 8th International Conference on Information Technology (ICIT), pp. 685-690.

[59] Mocanu, M., Cristea, V., Negru, C., Pop, F., Ciobanu, V., Dobre, C. (2015). Cloud-based architecture for farm management. In 2015 20th International Conference on Control Systems and Computer Science, pp. 814-819. https://doi.org/10.1109/CSCS.2015.55

[60] Chang, K.D., Chen, C.Y., Chen, J.L., Chao, H.C. (2011). Internet of things and cloud computing for future internet. In International Conference on Security-Enriched Urban Computing and Smart Grid, pp. 1-10. https://doi.org/10.1007/978-3-642-23948-9_1

[61] Giang, N.K., Blackstock, M., Lea, R., Leung, V.C. (2015). Developing IoT applications in the fog: A distributed dataflow approach. In 2015 5th International Conference on the Internet of Things (IOT), pp. 155-162. https://doi.org/10.1109/IOT.2015.7356560

[62] Shukla, A.J., Panchal, M.V., Patel, M.S. (2015). Intelligent greenhouse design based on internet of things (IoT). International Journal of Emerging Trends in Electrical and Electronics, 11(2): 78-86.

[63] Collina, M., Bartolucci, M., Vanelli-Coralli, A., Corazza, G.E. (2014). Internet of Things application layer protocol analysis over error and delay prone links. In 2014 7th Advanced Satellite Multimedia Systems Conference and the 13th Signal Processing for Space Communications Workshop (ASMS/SPSC), pp. 398-404. 

[64] Kelly, S.D.T., Suryadevara, N.K., Mukhopadhyay, S.C. (2013). Towards the implementation of IoT for environmental condition monitoring in homes. IEEE Sensors Journal, 13(10): 3846-3853. https://doi.org/10.1109/JSEN.2013.2263379

[65] Elijah, O., Rahman, T.A., Orikumhi, I., Leow, C.Y., Hindia, M.N. (2018). An overview of Internet of Things (IoT) and data analytics in agriculture: Benefits and challenges. IEEE Internet of Things Journal, 5(5): 3758-3773. 

[66] Fernández, J.E., Alcon, F., Diaz-Espejo, A., Hernandez-Santana, V., Cuevas, M.V. (2020). Water use indicators and economic analysis for on-farm irrigation decision: A case study of a super high density olive tree orchard. Agricultural Water Management, 237: 106074.