Behavioural sensor data as randomness source for IoT devices

Dinca, Lavinia Mihaela; Hancke, Gerhard P.

doi:10.1109/isie.2017.8001568

Cited by 10 publications

(9 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[22] observed far less entropy than expected in their experiments on entropy harvesting from physical sensors, and argued that the amount of min-entropy reported in [38] was overestimated. Moreover, L. Dinca and G. Hancke [21] studied using smart-phone sensors as entropy sources for random number generation during human gait, and concluded that the raw data collected from smart-phone sensors on the subjects during randomly walking on the city streets is highly predictable. The authors also argued that using data from any combination of two sensors can only slightly improve randomness.…”

Section: Discussionmentioning

confidence: 99%

“…The same rule is also applied to gyroscope signals, resulting two mixed binary sequences for each gait cycle group. Finally, the two binary sequences are bitwise-XOR operated, as combining two entropy sources can improve randomness [21], to produce the final (3 × C × L)-bit random binary sequence, which can be used as it is or be converted into decimals (random numbers) depending on the requirements of the applications.…”

Section: Re-indexing and Mixingmentioning

confidence: 99%

“…However, there are a few issues that have not been fully addressed. First, the raw sensor data collected during daily activities, such as walking [21], is periodic and predictable, thus, the random numbers generated from the raw sensor data are not truly ran-dom. Second, studies in [18,19,22] show entropy of a single sensor data is too low to be used as an entropy source for random number generation.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Random Number Generation Using Inertial Measurement Unit Signals for On-Body IoT Devices

Sun¹,

Lo²

2018

Living in the Internet of Things: Cybersecurity of the IoT - 2018

View full text Add to dashboard Cite

With increasing popularity of wearable and implantable technologies for medical applications, there is a growing concern on the security and data protection of the on-body Internet-of-Things (IoT) devices. As a solution, cryptographic system is often adopted to encrypt the data, and Random Number Generator (RNG) is of vital importance to such system. This paper proposes a new random number generation method for securing on-body IoT devices based on temporal signal variations of the outputs of the Inertial Measurement Units (IMU) worn by the users while walking. As most new wearable and implantable devices have built-in IMUs and walking gait signals can be extracted from these body sensors, this method can be applied and integrated into the cryptographic systems of these new devices. To generate the random numbers, this method divides IMU signals into gait cycles and generates bits by comparing energy differences between the sensor signals in a gait cycle and the averaged IMU signals in multiple gait cycles. The generated bits are then re-indexed in descending order by the absolute values of the associated energy differences to further randomise the data and generate high-entropy random numbers. Two datasets were used in the studies to generate random numbers, where were rigorously tested and passed four well-known randomness test suites, namely NIST-STS, ENT, Dieharder, and RaBiGeTe.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Re-indexing and Mixingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Random Number Generation Using Inertial Measurement Unit Signals for On-Body IoT Devices

Sun¹,

Lo²

2018

Living in the Internet of Things: Cybersecurity of the IoT - 2018

View full text Add to dashboard Cite

show abstract

“…It is the sequence of numbers used to generate random values. If the entropy or seed can be predicted then the PRNG fails [52]. In this regard, a lot of researchers ventured into the development of a lightweight random number generator that fits in this resource-constrained environment.…”

Section: Related Workmentioning

confidence: 99%

Modified Key Derivation Function for Enhanced Security of Speck in Resource-Constrained Internet of Things

Lustro¹

2021

IJCNIS

View full text Add to dashboard Cite

Randomness is an imperative component in every cryptographic algorithm to guarantee secret keys are unpredictable and secured against all forms of attacks. Speck generated sequence is non-random, a clear sign that it falls beyond the acceptable success rate when tested in statistical analysis. Thus, this study resolves the non-randomness by integrating a novel key derivation function that uses elementary operators designed for lightweight application. This design aims not to compromise performance when implemented on software and hardware. As a result, the modified Speck successfully passed the NIST SP 800 - 22 and Dieharder v3.31.0 Statistical Test Analysis as no p-value is flagged as failed during testing. Hence, making modified Speck cryptographically secured. Nevertheless, a 1.06% decrease in the figure of merit of the modified Speck still makes it worthier in a resource-constrained Internet of Things application as contrasted to Speck because it is proven to be beyond cryptographically secured.

show abstract

“…In addition, they designed a TRNG by combining the collected sensor values through a lightweight mixing algorithm. In 2017, Dinca et al analyzed the randomness of biometric data collected from 6 smartphone sensors: accelerometer, gyroscope, and linear acceleration, gravity, rotation and sound sensors [17]. They showed that since human gait is predictable, it should not be used as a random source.…”

Section: ) Random Number Generators Using Sensorsmentioning

confidence: 99%

Random Number Generator Using Sensors for Drone

2020

View full text Add to dashboard Cite

In recent years, drones have been widely used in many areas such as farming, movie making, surveillance and delivery. So, there is a need to protect these drones against security attacks including hijacking, spying and theft of stored data through the utilization of security mechanisms. Cryptographic keys are needed to operate these security mechanisms, and they must be generated by using random number generators which create unpredictable and non-regenerable random numbers. However, existing random number generators used in drones are not tailored for drones specifically as they use random sources generated on a desktop, not a drone. Recently, random number generators utilizing sensors in mobile phones and IoT (Internet of Things) devices have been studied, but are not appropriate for drones. In this paper, considering that drone sensors must be applied to flight and stationary modes, we proposed a drone specific random number generator called DroneRNG and implemented it. Then, we showed that our DroneRNG passed all of the NIST randomization tests and possesses better statistical properties and unpredictability than random number generators that are currently used in drones.

show abstract

Behavioural sensor data as randomness source for IoT devices

Cited by 10 publications

References 20 publications

Random Number Generation Using Inertial Measurement Unit Signals for On-Body IoT Devices

Random Number Generation Using Inertial Measurement Unit Signals for On-Body IoT Devices

Modified Key Derivation Function for Enhanced Security of Speck in Resource-Constrained Internet of Things

Random Number Generator Using Sensors for Drone

Contact Info

Product

Resources

About