Development of an Electronic Smart Safe Box Using Private Blockchain Technology

Abstract

Effective security has become a vital part of our daily routine, for example securing cash, jewelry, documents, and other valuables. One way is to protect these items by keeping them in a safe box. However, current safe boxes are vulnerable to a number of security attacks such as physical, repudiation, anonymously modifying, etc. In this regard, this paper presents a robust smart safe box using the blockchain solution to ensure each transaction and secure it from unauthorized access. Also, it traces back the local safe items by reading back to blockchain data transactions and protecting them from thieves by tracking the burglar with a real-time tracking location Global Positioning System (GPS) tracker and capturing a picture and video recording through the external camera. Similarly, with more security enhancement using the TensorFlow framework to analyze optical detection and identify things inside the box using an internal camera. At the same time, each transaction is stored by using Write Once and Read Many (WORM) patterns saved in local private blockchain data block logs. Finally, it uses three-way authentication mechanisms to unlock the safe box that authenticates the users, (1) face recognition, (2) fingerprint and (3) the keypad password. The performance of the proposed smart safe box is tested by a number of hardware and blockchain tests. The results reveal that the proposed smart safe box ensures safe and authenticated access to the box and records each transaction using blockchain technology. Further, the performance evaluation of the private blockchain significantly improved the integrity and safeness of the proposed smart safe box.

1. Introduction

The need to secure valuable things in a safe and secure traditional safe box at home with the use of different family members might lead to missing important items and documents. For example, a family member who keeps sensitive documents in the safe box and tries to find them may sometimes have difficulty remembering where they are and who took them or the last time the box was used. It refers to the nature of the human being, prone to forgetfulness, who might not remember who added or removed items from the safe box or when they were removed. However, with the traditional safe box approaches of using a key or keypad to safeguard valuable things, security is not guaranteed because the keypad password method can be easily seen by others using shoulder surfing attacks and that makes this security method useless. A smart safe is an integrated lock that would offer maximum efficiency in securing important materials from getting into unauthorized hands. In order to protect a safe box from various security attacks, a number of authentication mechanisms such as face recognition, fingerprint, and keypad password could be used. However, a smart safe used by a number of family members further requires a blockchain solution for securing and blocking unauthorized access.

A smart security box would work seamlessly with the blockchain as it provides secured data that only the owner can access [1]. A smart security box would work with blockchain allowing the placing of data in an open network securely. Additionally, it is secure and trusted as they are stored linearly following the chronology, making it difficult to go back and alter the contents of a block. Blockchain is by nature decentralized; it has a high level of transparency. Furthermore, this research work increases the traceability of data shared across a business network and saves costs with new efficiencies. A smart security box is a hardware wallet that gives additional security to information and funds. Furthermore, blockchain would offer immutability, allowing the saving of recorded data in the safe box without it being vulnerable to changes.

Individual authentication might fail. However, if the stated access options fail, the further security modules of the smart safe box are activated [2]. In that case, this smart safe box would have the installed security strategies described below:

• Motion detector and camera recorder: The camera is activated to detect recorded data when someone moves in front of the box.
• GPS tracker and alarm: If anyone attempts to take the box by force, the administrator can trace them with the GPS tracker.
• Optical detector: Used to detect things in the security box and collect them in a log file to find missing items.
• Private blockchain: Used to secure all private block transactions logged so that no one can modify them.

Conclusively, this research employs the knowledge of security boxes to keep logs safe by following the blockchain WORM ledger with limited permission for users to read. In the WORM ledger once the information on any transaction is written, it cannot be modified. This write protection assures the data cannot be tampered with and thus the integrity of the data cannot be compromised by any subject using the safe box. Additionally, tracing and tracking of the security box are enhanced since the security box would generate and send a Uniform Resource Locator (URL) with a GPS location and secure logs with blockchain if the box is stolen. Further, the use of blockchain technology is used to enhance the security of smart security boxes with the use of the Internet of Things (IoT) to detect local items and improve the detection to understand the locally stored things of value in different places. This paper has main contributions as listed below:

• To the best of our knowledge, we are the first to consider using a private blockchain solution in a smart safe box to enhance security and tracing things.
• This paper demonstrates to detect the listed most common local use items stored in the local traditional safe box to enhance the security of things instead of memorizing the things that lead to forgetfulness or skepticism.
• We enhanced the safe box to be more robust, safe, and secure with more protection against being robbed of the stored things of value that improve the performance for detecting the current items or to go back historically to when it was taken and when it was there which led to more robustness and security in storing things in the box.

The rest of the paper is divided into the following sections. A thorough and relevant discussion of the current schemes is presented in Section 2. The proposed scheme and implementation of the system are discussed in Section 3 followed by a performance evaluation of the security techniques in Section 4. Finally, the paper is concluded in Section 5.

2. Literature Review

In this era of the IoT, connectivity and network communication are largely exposed to a number of security attacks. Similarly, smart home devices such as safe boxes need robust security solutions based on blockchain and other relevant solutions [3,4,5]. In this regard, several previous researchers have presented several smart safe box securities and IoT mechanisms [6,7].

In a research work presented in [8], the authors developed a smart safe prototype based on IoT architecture with face and fingerprint recognition. They used specifically some features like personal identification number (PIN) code, two-way verification, and face recognition fingerprint. Due to limited security options, the safe box is vulnerable to various security attacks such as shoulder, physical, etc. In order to provide a guard against a physical attack, the authors propose a system as presented in [9] using a remote tracking security box with a multi-factor authentication system. The designed security box can be tracked remotely. In addition, they implement a biometric fingerprint sensor and notify when the security box arrives at the base station.

The work presented in [10] represents an automatic safe deposit box security system using Arduino Uno. A safe box with a security system and a microcontroller using the Arduino Uno contains features such as a buzzer, password keypad, and a fingerprint. They also thoroughly list the advantages and disadvantages of fingerprint mechanisms. However, the safe box is vulnerable to relocating attacks as the system does not provide a GPS tracking system. Similar research is performed in [11] to overcome the limitation in the article [10], by adding a global system for mobile communication (GSM) module to enhance the smart safe box’s security. Any attempt to relocate the smart safe box activates the GSM module which sends an SMS message to the authorized user. This works as a successfully executed SMS-based system in the smart box. This potential smart safe box can be used for personal and commercial purposes. However, this research relies on the availability of the internet. Thus, if there is no internet, the user cannot access the device.

The LockerSwarm [12] is a smart locker solution that can be used as a public locker in universities. These public lockers use the Arduino to design a smart lock that is used with a control of the electric door’s lock. Also, they use a Quick Response (QR) code that allows users to manage their lockers through the messaging app. The users can also share access to the locker with other colleagues or friends. Further, the LockerSwarm allows a system administrator to monitor and control all safes. Following the approach presented in [12], a survey of various types of smart lockers is presented in [13]. The authors identified a number of problems in designing a smart locker system and the possible solutions to overcome these problems. The authors also classified a number of areas where the smart locker can help in keeping information and documents safe from any possible type of security attack. However, the smart locker system presented in the survey is mainly related to how to avoid physical attacks.

Face recognition and One Time Password (OTP) are proposed in [14,15], where a locker system is installed with face recognition using an OTP service which is sent to the user via SMS and GPS. Typing the locker PIN is allowed with three attempts. However, this research does not use fingerprints to give access to the user. Also, the transaction to the smart safe box is not recorded which may reduce trust among the users of the safe box.

In order to record the transactions for safety purposes, blockchain technology has recently been introduced [16,17,18]. Similarly, blockchain technology has recently been researched for providing security to IoT systems. For instance, in [19], the authors presented the application of blockchain technology to ensure secure transactions to IoT devices and block the tempering of sensitive data using the Hyperledger Fabric blockchain. A scenario is implemented by IoT devices connected to the centralized server for sharing data where each transaction is recorded by implementing a block. Every participating peer keeps a copy of the ledger, and they also check and sustain the blockchain. Likewise, a system is proposed in [20] to ensure a decentralized mechanism to block the cyber security attacks on the system proposed in [19]. This system further enhances the security of the system proposed in [19] by providing a solution for the constrained devices to join the blockchain network efficiently and securely.

A blockchain is mainly based on preparing the hash of data and then linking it to each other as a block. Therefore, using this technology always adds many challenges such as requiring high processing systems, high memory, etc. In this regard, a system based on a lightweight consensus algorithm is introduced in [21]. The proposed lightweight blockchain technology efficiently meets the processing requirements required for IoT devices. Therefore, this technology could be an appropriate solution to use in IoT environments. However, as the IoT consists of many different events, employing such technology at multiple phases can increase the speed requirements.

Another blockchain solution is presented in [22] where the author introduces a transfer and store concept to IoT devices. Further, this study presents various challenges that can come across in implementing blockchain technology for IoT devices. The authors also mentioned using the Hyperledger Fabric blockchain as stated in [19]. However, the performance evaluation is performed in smart home environments where the requirement of speed is already provided via high-end systems. Therefore, using such a technology for the IoT environments constrained to limited speed may decrease the availability and suitability. In order to avoid the limitation of the system proposed in [22], the authors of the system presented in [23] come up with a system that introduces the concept of smart contract security. The idea of smart contract security is to secure the transaction for IoT products accessed by a group of users such as smart safe boxes in homes. However, the proposed system, which is based on Ethereum smart contracts, poses many challenges to the implement it in IoT environments.

As the IoT-based smart homes and cities are rapidly adding new things and, therefore, the security of accessing these things requires sophisticated and lightweight mechanisms. In this regard, a Proof of Concept (PoC)-based blockchain technology is introduced to handle the authorization issue to the things available in smart homes. According to the authors, a smart contract-based solution is implemented to ensure the integrity of the system. As this concept is purely designed for administration access, the practical implementation of such a system requires great care and tunning of a wide number of variables. One of the main reasons the authors suggested the use of blockchain is it provides transparency which in turn helps in auditability [24,25,26,27,28].

Concluding the above literature, it is worth mentioning that so far none of the blockchain technology can work efficiently for protecting smart safe box transactions. We have therefore proposed a system to handle the transactions in a smart safe box using an immutable blockchain technology as stated in Section 3.1. Also,

Table 1. A comparative analysis of the existing schemes.

	FR	FP	KP	Alarm	LCD	APP	GPS	RFID	IRIS	OTP	HS	SMS	PIR	Speaker
Six Tier Multipurpose Security Locker System Based on Arduino [1]		✓	✓			✓	✓					✓	✓
Smart Safe Prototype Based Internet of Things IoT [3]	✓			✓	✓	✓
3-Level Authentication for Bank Locker Security [8]	✓	✓		✓					✓	✓
Smart Locker: IoT based Intelligent Locker [29]	✓		✓		✓	✓				✓	✓
Development of a Remote Tracking Security Box [10]		✓	✓		✓		✓					✓
Design and Implementation of Bank Locker Security System [30]		✓				✓		✓
Advanced locker security system [31]			✓	✓	✓			✓			✓	✓
Bank Locker Security System [32]			✓	✓				✓				✓
Smart Bank Locker Security System [33]		✓	✓	✓	✓			✓				✓
This Research	✓	✓	✓	✓	✓	✓	✓			✓		✓	✓	✓

shows the most related research works relevant to our proposed work. All these research articles are mainly about the smart safe box, to discover how they are use and what enhancement features there are in the perspective of the IoT. The smart safe box is typically used to store precious stuff belonging to a person or group, and can use any available technology features to protect it such as a fingerprint, face recognition, alarm system, and tracking by GPS. The details of each critical used feature in this research are described in the proposed system section. The listed literature in Table 1 shows the essential hardware feature used in the below literature that introduced several critical points about the work carried out in this paper.

3. Materials and Methods

Blockchain technology is categorized into two categories: (1) public and (2) private. This research focused on the private blockchain because the private blockchain allows the sensitive data only seen by selected permitted users and no need for mining as a public blockchain. In addition, a blockchain is a digital ledger of each block of transactions with duplicates and distributes across the entire network of computer systems known as nodes. Also, blockchain by its nature is a transformative technology with a digital form of record-keeping that organizes information added to the ledger into blocks or data groups. Further, the individual blocks are designed to hold a certain amount of data; thus, new blocks keep adding to the ledger continually, leading to the formation of a chain to structure the data as a linked list. Moreover, each block has its unique identifier, a hash algorithm that protects not only the information contained in a block but also the place of the block along the chain by identifying the previous block before it. Also, once the block data has been added to the chain in the private blockchain, it is encrypted using a standard hash algorithm, which becomes permanent and unchangeable. Communications between components are known as transactions. There are different transactions in the private blockchain-based smart safe box, each designed for a specific function. For example, a store transaction is generated by components to log data and an access transaction is generated by the user to access the smart safe box.

The implementation of blockchain in this research is to add the transactional chain logged into a database when the user with permission wants to see the data logged in the data logs. Also, the blockchain utilizes the standard encryption algorithm to make it more confidential in each transaction chain. Further, the blockchain helps secure data to be stored and thus can be applied in a smart safe box. Further, blockchain data have three aspects of data structure, data model and datastore, and a consensus mechanism as presented in [34,35,36,37]. Further, each block is saved in chronological order with a block header, policy header, and block body. The block body records all the transactions started from the motion sensor until the box is closed or ends with no motion. Also, block body logs contain time, date, picture, video, and transaction ID to protect the standard hash algorithm for each block transaction. Figure 1 shows how the private blockchain implementation design works and it starts with:

(1)

The user opens the box using the three-way methods (face recognition, fingerprint, or keypad password).

(2)

The user adds the items inside the box, and the optical detector detects these items and stores them in the block.

(3)

Collect the data.

(4)

Save the data in the private blockchain.

(5)

Use the logs option to access data if needed.

As shown in Figure 1, the private blockchain concept used in the implementation of the proposed smart safe box consists of three blocks i.e., (1) Genesis Block 1, (2) Block 2, and (3) Block 3. The working concept of each of these blocks is given below.

• GENESIS BLOCK 1

  ○

  Contain hash for itself data

  ○

  User entry log

  ○

  Items were detected

  ○

  Time

• BLOCK 2

  ○

  Contain hash for itself data

  ○

  Previous hash

  ○

  User entry log

  ○

  Items were detected

  ○

  Time

• BLOCK 3

  ○

  Contain hash for itself data

  ○

  Previous hash

  ○

  User entry log

  ○

  Items were detected

  ○

  Time

Figure 2 explains the flowchart used for the smart safe box system. Each transaction is saved in the blockchain ledger at each block logs transaction. At the first step the smart safe box is connected to the internet; in the motion sensor, if there is any human motion deducted by the Passive Infra-Red (PIR) sensor, then the external camera starts recording videos in the event of movement, so if the box is stolen, the system sends the location to Khaznti app to allow it to track it. Otherwise, the registered authorized user can use the three-way authentication to unlock the safe box. The working of the proposed blockchain mechanism is illustrated in the following Section 3.1.

3.1. Working of the Blockchain in the Proposed Scheme

Initializing the Blockchain

A log or a block must be immutable. It shouldn’t be changeable. In fact, when we add entries to a block, the block may break, which is a fraud. One of the essential pillars of blockchain is ensuring that the data in the block is not modified, and here’s how to ensure that. When a block is mined and added to the blockchain with our log, a hash for the block is generated by combining the data (login, items detected) in the block. In Figure 1, the first block is the genesis block, and it has no previous hash by default due to the lack of the previous block. Instead, it has the hash, generated by combining its data and id, along with a calculated and uniform nonce. The next block has the same data field as the original block, except that the previous hash will not be empty. Instead, it takes the current hash value from the previous block and assigns the value to the previous hash variable in the current block. Further, current and previous blocks are linked based on the current hash and the previous hash field in the block. Note that if the current hash of the previous block changes, the previous hash field of the current block will not change, resulting in an error in the blockchain.

In a blockchain, we need a unique identifier for each block, such as a fingerprint. The way to ensure this is through an algorithm developed by the National Security Agency (NSA), Secure Hash Algorithm 256 (SHA256). In addition, in blockchain, data from the block and previous hash and other attributes are used to generate the block hash, which serves as a unique fingerprint for the block.

Similarly, the user will use three-way security authentication to access the smart safe box, starting with face recognition, fingerprint, or entering the password using the keypad. In addition, if it is a verified user, the system will unlock the box. However, if the user makes three wrong attempts to enter the system, the system produces an alarm and contacts the administrator.

3.2. Hardware Components

Figure 3 shows the proposed design hardware structure of the smart safe box as it consists of a built-in GPS sensor and camera in front of it for face recognition and records, a keypad to enter the PIN code, and a Liquid Crystal Display (LCD) to show the smart safe box status. For more security there is also a fingerprint sensor and alert unit to produce an alarm sound if there is any wrong attempt to open a smart safe box. Also, the proposed system has an internal optical detector to detect the contents of the box.

3.2.1. Three-Way Security Authentication

Some works of literature use further security verification. This research applies three-way security authentication. First is the face recognition if it captures the photo of the user, then stores it in the logs and authenticates it with the same verified user, then it will unlock the safe box. If it is not the same authenticated user it will store the snapped picture and will continue recording a video, then the expected user can move to fingerprint authenticator which is the second step in the three-way security authentication that allows the user to scan the finger. If it is verified with the same stored user, it will unlock the device. Otherwise, it will be additional evidence that the user is trying to get into the safe box. Finally, in the third step, using a keypad to enter the pin after recording the picture and video with the fingerprint: all of these can be taken as evidence. Finally, if the users exceed the time limit and do not close the door it will alert the alarm and send a notification to the administrator to investigate after it gives a three-count time to allow the mistakes or errors; if not it will alert an administrator to take possible countermeasures.

3.2.2. Cameras

We have two cameras, one is an external camera and the other one is an internal camera. Each one has its use and is explained below.

External Camera

The external camera is used to detect face recognition to recognize users’ faces to authenticate or compare with approved verified users as registered in the system.

Face or Facial Recognition (FR)

The FR is a method that is used to authenticate the human face from the external camera. Also, it authenticates the user’s face by calling the facial saved image data of the user for the purpose to authenticate or unlock the device by using a camera as stated in [38].

Internal Camera

The internal camera inside the box provides an optical detector to detect the items inside the safe box and log them inside the Khaznti system. Further, the user has access to the logs section and can check if any item was missing.

3.2.3. Fingerprint (FP)

This is used to validate the human finger because each human has a unique fingerprint divided into three shape types; loops, whorls, and arches. The arches are normally used after it is recognized and it sets the user data by scanning the user fingerprint. Nowadays, the new type of mobile and computer device uses fingerprints to unlock the device to do mobile transactions to authenticate the users.

3.2.4. Keypad Password (KP)

The traditional keypad password has some security issues for example, the pin might be guessed. Also, a thermal sensor or thermal camera can be used to steal the user’s password. Also, the attacker might use a shoulder surfing attack. Because of the previous issues, this research enhances the security, and it maintains the last step to use the keypad password, then it will alert the administrator to let them take appropriate action.

3.2.5. Arduino UNO

Arduino UNO is an open-source microcontroller board used to control all of the sensors and modules. This research used an Arduino Uno microcontroller.

3.2.6. Buzzer/Alarm

The alarm or buzzer is used to produce an alarm sound if there is any attempt to open the box or what program for it.

3.2.7. Battery

The proposed system is powered using a power source, so the first risk is when the electricity is cut down. The system is powered using a built-in battery. If the smart safe box is stolen, the administrator can trace the burglar using the GPS Tracker.

3.2.8. LCD Screen

The LCD Screen welcomes the user and shows time and date as well as the most important information of the user and writes the deducted items.

3.2.9. GPS Tracker

By using the GPS tracker system, users can track the box in the GPS tracking system offering the safe box’s current location at any given time and blockchain for data logs for more security. The GPS module needs an internet location to update the security box location. This means that the system is provided with a GSM SIM card to let the administrator see the updated location.

3.2.10. Speaker

The use of the speaker to run the voice of the stored recognized items like passport, cash, Swara, Baqama, etc. The administrator can speak to whoever is next to the safe box. Also, it has an announcing alerting voice.

3.2.11. PIR Sensor

The PIR or motion sensor is usually used in an external safe box and connected to the external camera to start recording the video and detect any movement in front of the box and store it in the blockchain database.

3.2.12. RFID

The main issue with using an RFID card is anyone who holds the card can access the device [39]. However, anyone who steals or grabs the card has full access to open the safe security box for those who are unauthorized users and without real user knowledge. Further, the RFID, IRIS authentication, and heat sensor (HS) are not used in this research. The IRIS authentication is a biometric identification system. This technique can take an image of the eyes as each person has unique eyes [40], therefore, this research will not use an IRIS sensor to unlock the device because there is already a reliable alternative solution. The heat sensor detects the box’s heat to identify unauthorized access by using heat generated to break the box [41]. In addition, the researchers decided to use a motion sensor and recording camera to enhance the result instead of using a heat sensor that has limited use.

3.3. Software Components

The proposed research helps people to protect a valuable thing, not in the usual way but with IoT technology by using different ways to protect such as face and fingerprint recognition. In addition, it can also help in tracking the box in case it is stolen. This comes with the web application to control everything such as creating user changes and access to logs files to see who has opened the box.

3.3.1. Optical Detector and TensorFlow

TensorFlow is an open-source machine learning from Google [40]. The proposed system uses a detection system that can detect any custom object by adding classified images [40]. The steps of the optical detector are generally divided into three i.e., (1) collect images from the video shot from the internal camera of each item or thing and create a data set to label the categories of the item’s picture with the item category, (2) analysis and feature recognition of the things by the camera, and (3) collect the data from the camera and sort it into a logs file to transfer to the blockchain system. We listed the most common items that are stored in the local safe by creating our custom module of these items like cash, jewelry watches, and other objects and documents. The jewelry is either silver or gold with diamonds or gemstones and is normally worn by the girls and women of a family.

3.3.2. Khaznti App

The name “Khazna” means a metal box with a locker key to keep money and store valuable items, and in the Arabic accent, a letter “ti” is added to the name to become the property of the person. By using Khaznti app the proposed application provides three-way security authentication and local blockchain solution and an optical detector. For instance, a person in the family adds the passport to the safe box, and another family member opens it to add the watch to it. However, the second person grabs the passport and did not notify the first person about it. However, using the proposed system the first person can check his missing items. Also, the administrator can manage the users by adding or removing, modifying profiles, tracking the box, access to the logs, and switching on or off the alarm system.

3.3.3. One-Time Password (OTP)

An OTP is a dynamic password for using one time in a period [41]. In addition, a user can send a request to open the box, and the box generates one password for one-time use only. This research used an OTP which is required after the keypad pin to verify the user access.

3.3.4. SMS/GSM

By using SMS messages can be sent and received like OTP numbers sending alerts to administrators connected with the Global System for GSM [42]. It is used in the proposed system to track the box if someone has stolen it. Similarly, the privacy and security of data in the case the safe box is stolen is important and therefore, requires privacy and data security mechanism to ensure the integrity of the safe box‘s contents [43].

4. Results

The proposed system is tested against a number of hardware and blockchain tests. These tests are listed below.

4.1. Hardware Tests

4.1.1. Face Recognition Test

This is to test whether the system can recognize faces that are already registered and faces that are not registered. We tested some of the sample’s faces registered and non-registered. In order to test face recognition, 40 subjects were registered with the smart safe box. These subjects are then tested using the face recognition system of the smart safe box with different noise densities. These noise densities were selected by covering the face with various things such as face masks, color, hijab, etc. The accuracy of recognizing the tests is shown in the following Figure 3.

As we can see, with the noise density of 0.4, the recognition rate was more than 90%. This indicates that the proposed face recognition system is strong enough to handle various types of attacks such as covering the face with face masks, etc. It is worth mentioning that the face recognition system is used to identify and block any unauthorized access to the smart safe box. It is, therefore, not designed to compare the proposed recognition system to compare it with other existing techniques.

4.1.2. Keypad Test

The purpose of this is to test whether the system can recognize keypad passwords that are already registered in the system and if a password is entered wrongly three times, it may result in locking the system for a particular person or time. In order to test the system, a few of the correct and incorrect passwords were tested on the system. These results are shown in the following

Table 2. Keypad test results with correct and incorrect passwords.

Password	Number of Tries	Result
9537	one	Recognize
7362	one	Unrecognize
1838	3 times	Unrecognize and locked the box

.

The results from the above test are also shown in the following Figure 4 and Figure 5. The smart safe box turned on the green and red color when the password is entered correctly and incorrectly, respectively.

4.1.3. Fingerprint Test

This test was performed to determine the ability of the fingerprint sensor to scan the fingerprints. Fingerprints must be able to distinguish between fingerprints registered in Arduino and fingerprints that have not yet been registered. In this test, two fingers of the right hand are registered first for the subject. The system will not allow the subject if it applies an unregistered finger. The results from this test are shown in the following

Table 3. Testing of registered and unregistered fingerprints.

Finger	Status	Response
Left thumb	Registered	Recognized
Left Forefinger	Registered	Recognized
Left Middle Finger	Registered	Recognized
Left Finger ring	Register	Recognized
Left Little finger	Register	Recognized
Right Thumb	Not Register	Unrecognized
Right Forefinger	Not Register	Unrecognized
Right Middle Finger	Not Register	Unrecognized
Right Finger ring	Not Register	Unrecognized
Right Thumb	Not Register	Unrecognized

.

4.1.4. GPS Test

This test was performed to determine the ability of the smart box and whether it can send a warning to the administrator upon moving it without administrator permission. The proposed GPS must be able to send the data to the Arduino and the map to the app registered on each subject’s device. The results are shown in the following Figure 6.

As seen in Figure 6, the smart safe box starts sending its location to the registered users by moving it from one place to another. These results also reveal that the integrity of the smart safe box is compromised by moving it from its current location.

4.2. Evaluation of Proposed Blockchain System for Smart Safe Box

To test the performance of the blockchain technology used to secure the transaction in the smart safe box, we considered three different nodes performing the various transaction at different times of the day. Initially, node1 performs the transaction in the smart safe box by adding document1 to it. The transaction is sent to the server connected to the smart safe box via internet access. After some time, node2 performs a transaction by adding a new document to the smart safe box, however, removing some part of the document1. The transaction is recorded and sent to the server. Finally, node1 accessed the smart safe box for document1, which was altered by node2. Node1 checks the transaction of the log stored in the server upon each transaction to the smart safe box. Finally, node1 found that the transaction performed by node2 alters document1 and thus a security breach is identified by checking the previous transactions. The following Figure 7 shows the transaction of the above scenario.

The concept of the proposed blockchain is based on the consensus algorithm as discussed in Section 3. To implement the concept of private blockchain in the proposed scheme, we have considered a node opening the smart box as a leader and the rest of the subjects as witnesses. To perform a transaction, a leader first generates a transaction message and sends it to the witnesses. Unlike the voting-based consensus algorithm, the leader is not randomly selected in our proposed private blockchain solution. However, a node that is interested in the transaction can become a leader and the rest of them a witness. As soon as a witness receives a transaction request from the leader, it starts collecting the signature of the rest of the witness nodes. After collecting the signatures from the rest of the witness nodes, each witness node marks the transaction as a valid transaction. Finally, the signed transaction request is broadcasted to each witness. To test the above-mentioned scenario in the proposed smart safe box, we considered a situation where a transaction is successful only and therefore ignored the situation where a transaction is unsuccessful. In the case of a successful transaction, all the witnesses sign the transaction, and the final copy is broadcast to every witness successfully. Further, we considered all the participating witness nodes connected to the same network. Also, two different results were calculated i.e., the variation in (1) overhead with different transactions per block (t/b) and (2) throughput with transaction per second (t/s) against a different number of nodes. We have noticed that as the non-consensus activities increase, the overheads also increase. It can be seen in Figure 8a, that 70% of the overheads are caused by non-consensus activities, and the remaining 30% are spent on consensus activities. However, as the nodes are connected to the same network, the throughput is high as shown in Figure 8b.

Concluding the practical implementation of the system, we have noticed that the system may be a suitable solution for those who are sharing a common place for securing their important contents. For instance, in a joint family account in a bank, mutual documents, etc. the goal is to protect data from individual access and thus require the authentication of each entity modifying and updating the important things such as jewelry, documents, etc.

5. Conclusions

In this paper, we presented the development of a smart safe box based on private blockchain technology and three-way authentication mechanisms. To the best of our knowledge and study, this is the first time that private blockchain technology has been introduced in the design of a smart security safe box. In addition, contrary to traditional safety boxes, the proposed smart safe box has fingerprint and face recognition along with the keypad pin, using three levels of security. Further, the Khaznti app provides efficient security and easy monitoring of the contents of the box. Finally, we tested the proposed smart safe box using several use cases of letting the users enter and remove important documents and items. The results reveal that the proposed smart safe box has enough security to mitigate the risks of misplacement and stealing important items and documents. Similarly, the addition of blockchain support ensures the approval of every concerned user in accessing and modifying the items of the smart safe box. In the future, we will plan to enhance the detection system to detect more accurate items such as scanning a document and can access it remotely and enhance the alert system to detect if intruders are touching or grabbing safe things like a family member grabbing another family member’s jewelry.