Joshua D. Griffin

Experimental Demonstration of Complex Image Theory and Application to Position Measurement

Measurements of the magnetoquasistatic fields generated from a magnetic dipole (an electrically small current loop) located above the Earth are presented and compared to calculations using complex image theory. With a horizontal (i.e., the surface normal parallel to the Earth) emitting loop located at a height of h and a copolarized horizontal receiving loop located at a height of z ≥ 0, coupling between the dipoles was measured for distances up to 50 m along a direction perpendicular to the surface normal axes of the loops. Inverting the theoretical expressions to estimate the distance from measured field values resulted in peak and rms distance estimation errors of 23.01 and 11.74 cm, respectively, for distances between 1.3 and 34.2 m. Received signals were not strongly affected by the proximity of a group of people even when the line of sight was obstructed.

Accurate Phase-Based Ranging Measurements for Backscatter RFID Tags

Short-distance, millimeter-level ranging based on backscatter RF tags remains a challenge today. We propose a composite dual-frequency continuous-wave and continuous-wave (DFCW/CW) radar system for localization of backscatter RF tags. The coarse DFCW range result is used to correct the cycle ambiguity of the more accurate CW range result. A minimum mean square error (MMSE)-based method is employed to combine the DFCW and CW ranging data, given spectrum constraints. The composite radar concept is demonstrated with a custom 5.8-GHz backscatter RF tag system in a typical indoor laboratory environment with a strong line-of-sight (LOS). Results show that the MMSE technique can resolve the CW cycle ambiguity for a moving tag and results in a range error of approximately 1 mm after an acquisition period.

RFID shakables: pairing radio-frequency identification tags with the help of gesture recognition

A novel approach for pairing RFID-enabled devices is introduced and evaluated in this work. Two or more devices are moved simultaneously through the radio field in close proximity of one or more RFID readers. Gesture recognition is applied to identify the movements of the devices, to mark them as a pair. This application is of interest for social networks and game applications in which play patterns with RFID-enabled toys are used to establish virtual friendships. In wireless networking, it can be used for user-friendly association of devices. The approach introduced here works with off-the-shelf passive RFID tags, as it is software-based and does not require hardware or protocol modifications. Every RFID reader constantly seeks for tags, thus, as soon as one tag is in its vicinity, the reader reports the presence of the tag. Such binary information is used to recognize the movement of tags and to pair them, if the gesture patterns match each other. We show via experimental evaluation that this feature can be easily implemented. We determine the required gesture interval duration and characteristics for accurate gesture and matching detection.

Two-dimensional position measurement using magnetoquasistatic fields

Two-dimensional (2-D) measurements of the magnetoquasistatic fields generated from a magnetic dipole (an electrically small current loop) located above the earth are compared to calculations using complex image theory. The magnetoquasistatic coupling between a vertical (i.e., surface normal parallel to the earth) emitting loop and seven vertical receiving loops was measured in a two-dimensional x−y grid of 27.43 m by 27.43 m, all above the earth, where the receiving loops were located outside this grid. Inverting the theoretical expressions to estimate two-dimensional position from measured field values resulted in an average geometric position error of 1.08 m (100th percentile of the measured grid), and an average error of 0.89 m for 95th percentile of measured grid.

Higher order loop corrections for short range magnetoquasistatic position tracking

Magnetoquasistatic position tracking has been shown to be an excellent technique to measure distances between an emitting and receiving loop for distances up to 50 m along a direction perpendicular to the surface normal of the loops [1]. For short distances from the emitting loop (i.e., less than about ten loop radii) there is an error in the estimated distance. In this paper, we examine the cause of this error and show that a significant portion is due to the simplification of the emitting loop as a simple magnetic dipole. By including a more accurate expression of the source field, errors can be significantly reduced. We show that the first correction term results in a reduction in rms and peak distance estimation error of 12.51 cm (54.44 %) and 11.27 cm (44.72 %), respectively, for distances less than 1.5δ, where δ is the skin depth.

A bend transducer for backscatter RFID sensors

Backscatter RFID sensors are beneficial for sensing applications where small, low-maintenance sensor nodes are needed. In this paper, a transducer to monitor the curvature of a tagged object for backscatter RFID senors operating at 5.8GHz is presented. It is shown, by vector network analyzer measurements, that the presented prototype, an open-circuited microstrip line resonator, changes its input impedance with bending and thus, proves feasible for integration in an RFID sensor. Analysis of the transducers performance when integrated in a backscatter sensor is provided.

Sensing human activities with resonant tuning

Designing new interactive experiences requires effective methods for sensing human activities. In this paper, we propose new sensor architecture based on tracking changes in the resonant frequency of objects with which users interact.

Experimental demonstration of complex image theory for vertical magnetic dipoles with applications to remote sensing and position tracking

Position and orientation tracking in the magnetoquasistatic region of an electrically small transmitting loop antenna make use of complex image theory (CIT), which is an approximate algebraic model of the fields above a lossy dielectric, semiinfinite half-space. Experimental demonstrations of CIT from a transmitting loop that approximates a horizontal magnetic dipole (HMD) have been reported previously. This work reports the first experimental demonstration of CIT from a vertical magnetic dipole (VMD) in remote sensing and position-tracking applications.

Experimental study on the effects of groups of people on magnetoquasistatic positioning accuracy

Position and orientation measurements have been demonstrated, recently, using low-frequency magnetoquasistatic fields and complex image theory for distances up to 50 m [1]. The key motivation for using magnetoquasistatic fields is to enable accurate estimation of an objects position and orientation when near weakly conducting dielectric obstacles, e.g., groups of people. An example application is tracking an American football during game-play [1]. In this paper, we present measurements using the magnetoquasistatic technique to show that the presence of a large group of 25 people introduces a peak distance error of less than 4.5 cm for an emitter-receiver distance of 10 m.

Error Reduction in Magnetoquasistatic Positioning Using Orthogonal Emitter Measurements

Measurements of the emitted magnetoquasistatic fields generated by a vertical emitting loop and detected at the terminals of seven fixed vertical receiving loops, all located above earth, are used to solve for position and orientation of the emitter. The coupling between the mobile emitting and fixed receiving loops was measured over a 3 × 3 emitter grid spanning an 18 ×18-m ² area and for azimuthal orientations between 0^° and 330^° at 30^° increments. Inverting the theoretical coupling expressions for two-dimensional position and azimuthal orientation resulted in a mean position and orientation error of 0.62 m and 2.86^°, respectively. Calculations including orthogonal-emitter configurations resulted in a mean position and orientation error of 0.21 m and 1.12^° , respectively, which represents a 66.1% and 60.8% reduction in error, respectively.