Table 2.

Comparison among different published sensor technologies for monitoring joints.

Ref.	Types of Sensor/Technology	Monitored Joint Parameters *	Measure	Method of Analysis	Advantages	Limitations
[25,26,27,28]	Optical fiber sensors	Angle	Attenuation of the transmitted optical signal power	Using the relation between the attenuation and the bending angle of the fiber	High resolution Flexibility Light-weight Long term reliability Immunity to electromagnetic interference	Limited measurement range (Angle) Nonlinearity Sensitive to temperature and humidity
[17,29,30]	Optical-based goniometer	Angle	Planar motion of an optical navigation sensor	Detecting navigation of the sensor using a miniature camera to calculate the bending of the joint	Compact and light-weight Flexibility High accuracy High speed of reaction	Sensitive to placement location May hinder natural joint movement during operation 3D sensing may not be possible
[31,32,33,34,35,36]	Imaging and video-based tracking system	Angle, motion, skeletal tracking	Visual data of several human actions	Skeletal tracking using anthropometric constraints and known joint locations in reference videos **	High accuracy and sensitivity Able to capture movements of multiple joints at a time No body-worn sensors are needed	Complex procedure with expensive infrastructure and sophisticated analyses Limited coverage area Requires body markers and adequate lighting condition for accurate measurements Unreliable to differentiate between near and far parts of human body, and for postures having self-occlusions ***
[37]	Textile-based conductive wire sensors	Angle	Changes of resistance	Changes of resistance are directly proportional to joint angles	Comfortable and suitable for long-term monitoring Simple mechanism One-time calibration Low-cost	Low resolution Low accuracy Nonlinearity Material uncertainties and hysteresis
[38,39,40,41,42]	Textile-based flex sensors	Angle	Changes of resistance	Changes of resistance are directly proportional to joint angles	Flexibility and stretchability Easily attachable with comfortable garments Low-cost	Fragile and lower lifetime (Prone to be damaged due to numerous bending) Low accuracy with noisy signal Nonlinearity Sensors are wide and affixing multiple sensors on the supportive garments is not feasible
[43,44,45,46,47]	Textile-based strain sensors	Angle, motion and rotation	Changes of resistance	Changes of resistance are directly proportional to joint angles and motion	Flexibility and stretchability High sensitivity Low-cost	Performance degradation due to large mechanical strains and rigorous deformations Signal drift due to the viscoelasticity of materials Limited to sense movements in the sagittal plane
[48]	Piezoresistive sensors – chopped carbon fiber (CCF)/polydimethylsiloxane (PDMS) yarns	Motion	Changes of resistance	Variation of relative resistance under mechanical deformation due to joint movements	Flexibility High sensitivity Easy integration into textile structures	Nonlinearity Material uncertainties and hysteresis Applying higher strain may cause piezoresistive performance (i.e., sensitivity) decay and delays the piezoresistivity transition
[49,50,51,52,53,54]	Smartphone sensors –accelerometer, gyroscope, magnetometer and camera	Angle, motion	Acceleration, inclination and camera measurements	Using smartphone applications to gather inbuilt sensors and camera data for measuring the range of motion	No external sensors are needed No external communication and data processing module are needed Applications are easy to implement	Lower accuracy comparing to other external sensors-based applications Difficult to place smartphones around different body joints Unable to monitor complex joint movements No standardized testing procedures are reported for clinical application
[55,56,57,58,59]	Acoustic emission (AE) sensors –piezoelectric-films/MEMS-based microphones	Angle, motion	High-frequency sound signal occurring during joint motion	Changes of surface resistance due to acoustic emission	Low-cost Light-weight Easy to attach around different body joints	High background and interface noise Nonlinearity Low accuracy
[60]	Gyroscope	Angle	Three axes angular rate	Joint angle is calculated by comparing the angular rate between two calibrated gyroscopes (below and above the joint)	Small size Low-cost Light-weight High resolution Easy to attach around different body joints	Produces some large drift over time Complex algorithms are needed to reduce noise and drift error At least two sensors are needed to measure accurate angle
[61]	Magnetometer	Angle, motion	Change of magnetic field	Change of magnetic field is directly proportional to joint motion	Feasible to measure complex joint angles Easy to control with digital circuits	Interference in the magnetic field by ferromagnetic and EMF-producing objects in the environment may decrease the accuracy of measurement Unreliable for detecting the orientations of joints in a 3D environment
[62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82]	Inertial measurement unit (IMU) sensors –accelerometer, gyroscope and magnetometer	Angle, motion, skeletal tracking	Three-dimensional acceleration, angular rate and the magnetic field vector	Three-dimensional angular velocities and linear accelerations are used to detect the position and orientation. Relative data from two calibrated IMUs are compared for tracking the joint angle and gait analysis	A combination of three sensors (Accelerometer, gyroscope and magnetometer) Compact and light-weight Small size Low-cost High resolution High accuracy Easy to attach around different body joints Built-in wireless module Built-in algorithms in new generation IMU sensors for calibration and to fix the sensors’ orientation with respect to a global fixed coordinate system Reliable for detecting the position and orientations of joints in a 3D environment	Sensors alignment is required in a multiple IMUs-based joint monitoring system Drift error from gyroscope (possible to compensate by fusing data from accelerometer and gyroscope)

* Joint angle: the angle between the two segments on either side of the joint; joint motion: the combination of the angle and the orientation of the joint; skeletal tracking: a technique to build a skeletal model of a human body by detecting the joint positions. ** Anthropometric constraints: size, shape and composition of the human body. *** Self-occlusion: one part of an object is occluded by another part from a certain viewpoint.