INTRODUCTION:

Agriculture is the major industry in Indian economy. 60% of the employed population contribute towards agriculture. As there is a rapid increase of population resulting various problems such as competition, exploitation of land, scarcity of food, shortage of water, and lack of natural resources. Consequently, rapid need of food resources becomes a huge challenge for agriculture sector. Food production, cleaning, storage, supply chain and preserving are various stages in agriculture. Growing requirements with a blend of organic farming practices and application of smart tools in agriculture guarantee a step towards sustainable development.

An optimal solution for critical trials in agriculture sector can be provided by emerging advancements of IT facilities more precisely the advent of IoT. Among various IoT devices invention of UAVs with a combination of image processing and artificial intelligence and data science technologies can form a networked solution for aforementioned challenges in order to gain precision agriculture. Such farming techniques improves quality crops with higher yields and low time consumption.

Numerous applications such as defence, medicine, marketing, farming, animal rearing and traffic monitoring are major developments of UAVs with enormous potential and smart networking. The developed drones are equipped with cameras, GPS receivers, sensors and services of IoT forming an advanced air borne network for communication [1].

Precision agriculture adapting UAVs include soil identification, vegetation health, aerial monitoring, irrigation, spraying fertilizers and pesticides, weeds removal and identification of failed crops using image analytics. Extraction of vegetation indices, monitoring variations in crop health, stress management, crop dis- eases and nutrient deficiencies can be captured by NDVI (Normalized Difference Vegetation Index) using image analytics.

Smart farming re-initiates organic practices in agriculture by sensor devices in a synergetic manner for improving crop yields and superiority grains. Although the industry has a rapid growth but heterogeneous internet-connected devices are ex- posed to cyber threats and vulnerabilities in precision agriculture. These attacks exploit drones by injecting unsafe malwares that corrupt the farming environment. For instance, the system instruction can be manipulated remotely by hackers like destroying entire fields, flooding farmlands, excess spraying of pesticides and fertilizers referred as agro-terrorism causing potential crop loss for nation.

Agro-terrorism has a significant impact on food security globally exploiting health of consumers. Cybercrimes at farms, food processing, transport and preserving areas requires a serious attention for developing a defensive food ecosystem infrastructure to protect from hackers and cyber criminals. Comprehensive research securing these networked devices for a smart food ecosystem is the current challenge in adapting precision agriculture.

Rapid growth of smart farming includes advanced farm equipment, labor sharing, prevailing environmental conditions influence operational decisions in agriculture. An attack can be an internal or external criminal so advances security models need to be implemented for a specific farming environment. The development of smart farm technologies therefore, demands further research before wide adoption in the community. Generating huge data from sensors and securing this information without any data leakage, impairment of physical devices, software, network and hardware equipment can persuade complex consequences [2].

Smart food ecosystem in the working model interacts with various devices. Sensor generate data from fields by random command instructions through IoT applications. The devices are connected via edges to a gateway enabling communications with in-farm equipment. Image analytics, timely agronomy analytics and simultaneously cloud storage in the data lakes are described in the figure 1. The data of soil moisture, water resources, pesticides sprays, energy management, nutrient monitoring and insects’ attacks are supervised using smart sensor machinery.

Fig. 1. Working Model of sensor-based UAVs application in Farming

Current advancement in smart farming demands for a tenacious security model for evolving domain. In this paper discussion on current threats, cyber-attacks, security research and cloud security models are the required challenges in precision farming. Key contributions of this study include to identify potential cyber-attacks in tech-based farming, impact of UAVs in rice production and OWASP based security model for preserving the data from leakage and hacking.

Section II discusses on security and privacy issues. Various Cyber-attacks on smart farming is been elaborated in section III. OWASP model for securing against the attacks is highlighted in section IV. Finally, section V draws the conclusion of the research paper.

Security and Privacy:

Smart farming practices has unleashed chances for orchestrate cyber-attacks. Hence, necessary measures need to be followed for improving the security and im- pairing vulnerabilities [3]. This section gives a detailed description of various privacy and security issues:

Prevention of data loss

Smart farming uses many numbers of sensor-based equipment that generates data. Loss or leakage of this collected data can occur either by an intruder or through illegal access. Anti-jamming devices can bypass all the security restrictions for an attacker. Crop health, pesticides spraying units, plant resources and water supply information can be intruded by hostile agents. Preserving the data from unauthorized access and increasing the security operations is the primary idea to ensure re- liable smart ecosystem.

Fig. 2. Working model of Edge computing in precision agriculture

Working model in figure 2 explains various security attacks through remote ac- cess then tracking IP address for accessing the smart devices. Data analytics is handled by third party service providers. These vendors can provide direct access to attackers by giving credentials. These potential cyber-attacks directly on the farm causing inoculating malicious software or SQL injections or by phishing.

Layer-wise security mechanisms:

Technological framework on IoT devices ensuring security through complex process focusing on application layer, network layer, linking layer and perception layer.

Application Layer security framework is described in figure 3. Enhancing the model by constraining the application with least privilege access controls and filtering the access configuration. Exporting the data into cloud by creating firewalls and IDS (Intrusion Detection Systems). Entry level scrutiny is examined by path identification and application of shallow network models.

Fig. 3. Application layer security mechanism

Network layer includes IP address security, transport security and secured rout- ing protocols in figure 4. IP Sec and ESP protocols for IP payload integrity, low- energy protocols for UAVs, IPv6 routing protocols for lossy networks, User Data- gram protocols for reliable nature of communication and several other services for Named Data Networking (NDN).

Fig. 4. Network Layer security components

In the linking layer end-end encryption, authentication bridges, validation models and network packet verification are channelized as security services. Multicast messaging, securing connection, executing architecture and trusted party authentication are monitored from sinkhole, routing attacks and wormhole attacks.

Fig. 5. Linking Layer security mechanisms

In perception layer includes protection of hardware, authorization and authentication services are discussed for an important security application [4]. Physical damage for the hardware components, symmetric key mechanisms for authentication, Oath protocols for authorization and digital signatures for end-to-end communication are various services described in figure 6.

Fig. 6. Perception Layer security components

Resolving Privacy issues:

Recent transformations in technology leads to loss of privacy for a human life. Everything is under surveillance and observation [5]. To balance these filters are applied for effective management. Pixelization, blurring, masking, steganography, morphing, warping, reversing and hiding are various filtering techniques associated with visual privacy. In particular, drone surveillance is the current prevailing issues in research.

Encryption techniques hide the data by algorithms, videos are scrambled by reversible techniques, morphing the images is demanded when there is an arose of privacy breaches. Patterns, numbering, connections and signatures are other privacy preserving schemes for centralized solution.

Cyber-attacks in Smart Farming:

This section evaluates various cyber-attacks in farming ecosystem. Data attacks, network attack and compliances are discussed in detailed in figure 7.

Data Attacks

Networking attacks

Compliance

• Internal data leakage

• Cloud data leakage

• False data injection attack

• Misinformation attack

• Malware injection

• Jamming attacks

• DoS

• Cyber terrorism

• Coud computing attacks

• Regulations

Fig. 7. Cyber-attacks for UAVs

Data Attacks:

· Internal Data leakage: An insider can become a threat by leaking the data. This could be done for money or due to personal grudge.

· Cloud data leakage: In smart farming all the data is recorded in cloud environment. This includes daily practises, economic and confidential information. As data is stored in a distributed and scattered data centres across the world. The laws of sensitive data localization were hence initiated to secure the data from cybercrimes [6].

· False data Injection: False data in measurements of humidity of soil moisture, usage of pesticides and fertilizers can cause a huge loss for a farmer this could damage the entire fields.

· Misinformation Attack: Guiding the farmers with misconceptions, false perceptions and faulty superstitious practices can damage the crop, outbreak of diseases in to the farm are few attacks causing loss for the farmer.

Networking Attacks:

· Malware Injection: Virus, trojan and trap door are several injections or com- mon attacks for connected smart equipment. Botnet, ZigBee and devastating effect can cause damage for the farm by stealing agricultural material and spoiling farm machinery.

· Jamming Attack: Jammers disrupt the farm equipment causing interruptions in collecting data and work inefficiently. Real time positioning of a UAV can be efficiently improved by enhancing the security and accessibility standards.

· DoS: A large distributed network that interrupt the services or cause a disruption to inter-connected nodes that results in abnormal function of the applications.

Compliance:

· Cyber Terrorism: Digital era in agriculture can attack a large group of people by food as a media. Such attacks are highly risk oriented and also ruin the nation’s reputation.

· Cloud Computing attacks: Cloud computing is the place where most of the invaders desire. They peep out for vulnerabilities if they could find a single virtual machine then spontaneously injecting the malwares takes place.

· Regulation: Many Countries and national food authorities enforces certain regulations and standards for ensuring the sustainable farm production. But smart farming techniques need advances protective measures to ensure data privacy and integrity.

FIBOR Model for Securing UAVs:

The proposed model acts as a countermeasure to safeguard UAVs against cyber- attacks. For each phase a typical method is adopted restricting the data loss, access and sharing by proper investigation techniques.

Fig. 8. FIBOR model for securing UAVs in smart farming

In the proposed FIBOR model described in figure 8 uses various applications for securing the smart farm. F- FANET, I- IDRF, B- BRUIDS, O-OWASP and R- REDLOCK algorithms are applied in the current system for preventing the data loss and unauthorized access.

FANET:

A safe recommendation for controlling the drones and the sensor generated in- formation using a hybrid IDS. Intrusion Detection System, microcontroller and processor are arranged in a Flying Adhoc Network (FANET). This network targets DoS and Distributed DoS attack on real-time traffic. This technique ensures effective and reliable framework for UAVs for various anomalies.

IDRF:

An UAV network can be secured using an Intrusion Detection and Response Framework (IDRF). This framework secures the networks from malicious threats, network attacks and data integrity. IDRF is a unique hybrid detection model ensuring multiple intrusions and cross suspicious nodes.

BRUIDS:

A rule-based technique for intrusion detection in drones is constructed using an adaptive Behaviour Rule-based UAV Intrusion Detection System (BRUIDS). This technique detects malicious airborne attacks and assess the survivability of drones. BRUIDS minimizes the detection errors, its adaptable model building reduces the false positives rate and optimizes the performance.

OWASP:

An Application Security Verification Standard (ASVP) provides technical security controls for a web application. Open Web Application Security Project (OWASP) performs a web application security verification standard to leverage the necessary security controls and a reliable environment. Many web application vulnerabilities such as XSS (Cross-Site Scripting), SQL Injection, broken authentication, insufficient logging, Insecure deserialization and misconfiguration are monitored to minimize these risks. Some of the requirements are given as below:

· Use as yardstick: A benchmark is set to measure the threat and trust for a web application by developers or the owners.

· Use as guide: Necessary regulations and guidelines are framed to build an adaptive security model by security control developers.

· Use during procurement: A security verification standards during part- nership between organizations.

REDLOCK:

A perfect cloud security platform is provided by REDLOCK. Security analytics, uninterruptable vulnerable identification, threat detection and compliance monitoring are significant features of REDLOCK platform. This network platform is able to perform threat forensics, sensing remediates worms and virus across cloud environment at the deployment of information. Resource optimization, user routines, network congestions, vulnerabilities in host and guest networks, third party authentication and threat intelligence are various Artificial intelligence based REDLOCK cloud network platforms.

Testing Farm parameters:

For the study conducted a rice fields are considered. The rice field for study is from Hoskote district, Karnataka State, India. The field size is of 20 acres and im- ages were taken from an altitude of 223 feet. Vegetation indices are collected by UAVs fetching the data of temperature, humidity, water and equipment sensors in- stalled in the fields.

Fig. 9. (a) Rice Fields taken for surveillance (b) Drone spraying the water for fields (c) Hyper spectral image for tracking the information of fields (d) Security of the fields by UAVs

Table 1: FIBOR model features for smart farming using UAVs

Components	Secured Solution	Protection
	FANET	End-to-end Encryption Physical security Strong Authentication Device Identification
	IDRF	Cryptography Hashing Firewalls IDS
	BRUIDS	Strong Accessibility methods Complex password policy Multi-level verification Physical Protection
	OWASP	Multi-Factor Authentication User Awareness Tested Application Updated Versions
	REDLOCK	Data confidentiality Data integrity User Privacy Non-Disclosure of Agreements
	IDS	Hoping Channeling Multiplexing Frequency Range

FIBOR model features are discussed in this above table 1. Model secures the UAVs from attacks and also protects the farm from many physical damages. This model improves the farm yield by 15% as shown in the following figure 10. As the amount of nutrients, supply of water, usage of pesticides, monitoring the temperature and humidity components measurements proves the usage of UAVs in an authentic way maximizes the quality and quantity of the crop.

In figure 10 various measures comparing the FIBOR model with traditional farming is elaborated. Amount of humidity in soil is uniformly maintained by UAVs as 28 throughout the crop season. This usually varies with respect to rainy season.

Amount of nutrient supply and pesticides usage is uniformly given for the farm through this the crop gives an optimized yield. With the application of secured FIBOR model using UAVs enhanced the crop yield by 15%.

Fig. 10. FIBOR model is compared with traditional farming techniques

CONCLUSION:

The enormous increase in the population and food requirements there is a current need for UAV enabled smart farming techniques. Many sectors defense, civil, health industries, marketing and agriculture are using the autonomous aviation vehicles for optimization of the necessities. However, in this digital era safety, privacy and security issues manifested with numerous cyber-attacks, vulnerabilities, risk and threats which are becoming a challenge to recommend the usage of UAVs in our activities. To ensure safer and secured usage of drones this article proposes a FIBOR model to prevent the UAVs from various attackers, hackers and cyber criminals. With the usage of drones 15% of yield has been increased in the rice fields used for the current research study.

REFERENCES:

1. G. Maanak, A. Mahmoud, K. Sajad and M. Sudip, "Security and Privacy in Smart Farming: Challenges and Opportunities," IEEE Access, pp. 1-22, 2020.

2. L. Thomas, A. Vasileios, B. Stamatia and S. Panagioti, "UAV IoT Framework Views and Challenges: Towards Protecting Drones as “Things”," MDPI - Sensors, pp. 4015-4036, 2018.

3. W. Sjaak, A. LanGe, V. Cor and B. Marc-Jeroen, "Big Data in Smart Farming - A review," Agricultural Systems - Elsevier, pp. 69-81, 2017.

4. S. Xiaojie, A. Xingshuang, Z. Qingxue, L. Huimin and X. Lianming, "State-of-the-Art Internet of Things in Protected Agriculture," Sensors - MDPI, pp. 1-24, 2019.

5. J. Molly, O. William L, T. Gregory F and D. Scott L, "Cyber Risk and Security Implications in Smart Agriculture and Food Systems," White Paper, pp. 1-20, 2019.

6. Y. Jean-Paul, N. Hassan, S. Ola and C. Ali, "Security analysis of drones systems: Attacks, limitations, and recommendations," Elsevier Public Health Emergency Collection, pp. 1- 22, 2020.

Received on 10.11.2020 Modified on 12.12.2020

Research Journal of Science and Technology. 2021; 13(2):119-126.

DOI: 10.52711/2349-2988.2021.00018