Imperial College London


Faculty of MedicineDepartment of Surgery & Cancer

Senior Lecturer



+44 (0)20 7594 0806benny.lo Website




B414BBessemer BuildingSouth Kensington Campus






BibTex format

author = {Sun, Y and Lo, B},
publisher = {IET},
title = {Random number generation using inertial measurement unit signals for on-body IoT devices},
url = {},

RIS format (EndNote, RefMan)

AB - With increasing popularity of wearable and implantable tech-nologies for medical applications, there is a growing concernon the security and data protection of the on-body Internet-of-Things (IoT) devices. As a solution, cryptographic system isoften adopted to encrypt the data, and Random Number Gen-erator (RNG) is of vital importance to such system. This paperproposes a new random number generation method for secur-ing on-body IoT devices based on temporal signal variationsof the outputs of the Inertial Measurement Units (IMU) wornby the users while walking. As most new wearable and im-plantable devices have built-in IMUs and walking gait signalscan be extracted from these body sensors, this method can beapplied and integrated into the cryptographic systems of thesenew devices. To generate the random numbers, this method di-vides IMU signals into gait cycles and generates bits by com-paring energy differences between the sensor signals in a gaitcycle and the averaged IMU signals in multiple gait cycles.The generated bits are then re-indexed in descending orderby the absolute values of the associated energy differences tofurther randomise the data and generate high-entropy randomnumbers. Two datasets were used in the studies to generaterandom numbers, where were rigorously tested and passed fourwell-known randomness test suites, namely NIST-STS, ENT,Dieharder, and RaBiGeTe.
AU - Sun,Y
AU - Lo,B
TI - Random number generation using inertial measurement unit signals for on-body IoT devices
UR -
ER -