I. INTRODUCTION
Usually, Radio Frequency IDentification (RFID) is a concept associated to the necessity of identifying a moving object or person passing by a specific position [Reference Finkenzeller1]. Think for instance to the systems currently in use in many highways to identify and charge the vehicles and to all the automatic doors that open when an RFID tag card is detected in the proximity of a purposely located transceiver. Further applications can be envisaged in the field of logistics and tracking goods, in the field of wireless sensor networks, where the sensors are embedded into the the RFID tag [Reference Pursula2, Reference Catarinucci, Colella and Tarricone3]. Their use is getting more and more widespread and pervasive going from logistic and healthcare to automotive applications.
Beyond these typical uses a new interest is coming out. In an increasing number of cases, a contact-less Self-Identification (SID) of a moving object is required to synchronize and set up actions aboard as soon as the vehicle crosses a certain point. Typical examples are robust road-to-car communications allowing for location-based systems (LBSs) information transfer (traffic jam, speed limits, alternative directions, and so on) to be directly logged into the car equipments [Reference Basat, Lim, Kim, Tentzeris and Laskar4, Reference Shaldover5].
In [Reference Roselli, Alimenti, Stopponi and Porzi6, Reference Alimenti, Mezzanotte, Palazzari, Placentino, Scarponi and Roselli7] an active Ku-band RFSID (RF Self IDentification) system was applied to Formula 1 (F1) race cars, where it has been used to reset the on-board electronics at each crossing of the finishing line as well as of specific points of interest along the path. It consists of a microwave receiver placed on a vehicle which has to be able to recognize when itself passes through a given virtual line. The virtual line is provided by a narrow electromagnetic beam incorporating relevant transmitted information (without lack of generality the information transmitted in the reference F1 application is limited only to the car identification code because additional information is forbidden by regulation). As a consequence of the identification process, a triggering pulse is generated and used to actuate whatever action required. The required robustness of such a system is achieved by working both on the digital and on the analog aspects: the transmitted signal is digitally modulated with gold sequences and a correlation receiver is adopted. Transmitting and receiving antennas, the performance of which strongly affect the time/space accuracy of the generated triggering pulse, have been designed accurately by means of a in-house Finite Difference Time Domain (FDTD) numerical simulator.
In [Reference Roselli, Palazzari, Alimenti and Mezzanotte8] the above system has been proposed for robust road-to-car communication. Thanks to a ±3° azimuthal lobe generated by the transmitting antenna and to a high data-rate digital modulation, in fact, the transmitted key sequence is detected by the receiver only at a precise space location and in a very short time, the latter being equal to about 75 µs. As a result, the transmission of location-dependent information as well as whatever location-based reaction of the identified moving object is actually possible. In addition, a high reliability is experienced during years of application in F1 races.
In this paper the results shown in [Reference Roselli, Palazzari, Alimenti and Mezzanotte8] have been extended showing the design of the RX patch antenna and going into details of the design, industrialization, and measurements of the entire system. The RX antenna, in fact, is a crucial component affecting the performance, the size, and thus the location of the receiver aboard the vehicle. A practical example of an industrialized folded version of the RFSID active tag will be shown. Such a tag, originally conceived for mounting in the very narrow rear-view mirror envelope of a F1 car can be thought as an optimized version of the original systems more suitable for critical mounting in practical moving objects (not only cars).
II. RFSID SYSTEM DESCRIPTION
The basic system is composed by a fixed transmitter, located in the path side, and by a receiver unit mounted on the moving object we want to alert; for example: a vehicle, as shown in Fig. 1, a good on a distribution line, an industrial machinery, etc. When the target crosses the zone illuminated by the transmitting antenna (shadowed area), the receiver unit senses a modulated microwave signal (see Section III). Once such a signal is detected, the receiver produces an output triggering pulse, the pulse duration corresponding to the time needed for the vehicle to travel from the point A to the point B. As a result, the apparatuses aboard can be synchronized exploiting the receiver output while, in the mean time, information is transferred from the fixed transmitter to the moving object.
Fig. 1. RF triggering of a fast moving object: principle of operation.
Although elementary, the model of Fig. 1 shows that the angular width of the illuminated zone is the main parameter determining the precision of the system. Formally, spatial and temporal uncertainties can be defined with respect to the ideal line OH (orthogonal to the path AB crossing the transmitter beam) as follows:
where R is the distance between the transmitter and the path AB and v is the speed of the moving object. For example, a temporal uncertainty of 250 ms for an object traveling 50 m far away from the transmitter at 20 m/s implies an angular width of the illuminated zone of about 12°.
It is important to note that, due to size and weight constraints always present in the tagged vehicles, goods, or machineries, the receiving antenna is necessarily small and thus of moderate directivity. As a consequence, the angular width θ is mainly associated to the performances of the transmitting antenna. This means that the transmitting antenna must be designed to have a narrow beam width in the azimuth plane for 2D horizontally moving objects.
Another transmitting antenna parameter that must be carefully accounted for at system level is the side-lobe to main-lobe ratio. Following the formulation introduced in [Reference Roselli, Alimenti, Stopponi and Porzi6], one obtains that the maximum range R max at which the active RFSID tag is able to detect the transmitter presence is given by
where EIRP = P TX·G TX is the effective power radiated by the transmitting antenna and S RX is the receiver sensitivity referred to the input. Equation (3) states that, for a fixed EIRP, high system ranges can be achieved only by reducing the minimum power S RX detectable by the receiver. Dealing with sensitive receiver has, however, a major drawback: when the distance R is small enough, the power associated to the transmitting antenna side lobes may be well beyond the S RX level, thus causing the information to be uploaded to the moving object in multiple angular directions. This results in a loss of accuracy in principle unacceptable. Defining PLR as the peak to side-lobe ratio of the transmitting antenna and following again [Reference Roselli, Alimenti, Stopponi and Porzi6], it is possible to evaluate the minimum distance R min to avoid the multiple information triggering:
From the ratio between (3) and (4) a general design criterion can be drawn:
If, for example, PLR = 25 dB, the ratio between maximum and minimum distances will be about 18. As a consequence, if a maximum range of 80 m is achieved by a suitable receiver design, the multiple triggering problem should not affect the system until the distance is reduced below about 5 m.
III. DIGITAL MODULATION
To transmit information from road to car, a digital FSK (Frequency Shift Keying) modulation scheme is adopted. The data rate is equal to 800 kbit/s allowing the key to be transferred in a very short time (hundred of microseconds). Without loss of generality, it is worth mentioning that, in this specific way, the system can work also in the case of high-speed vehicles.
When the RFSID system is used to identify the vehicle location, such as in unmanned vehicles or F1 race cars crossing the finishing line, a fixed code is transmitted. In this case it is important to:
– Reduce the interferences between closely spaced transmitters;
– Improve the reliability of the system in detecting a valid crossing;
– Allow for an accurate measure of the codeword arrival time (correlation receiver); and
– Select the destination moving object.
To achieve these goals a special class of pseudo-noise sequences, namely gold sequences, have been adopted [Reference Peterson9], due to their well-defined correlation properties. Given a code period of 31, a family of 33 gold codes can be generated. From theory, the auto-correlation of each sequence features a peak in the origin and low values in the other points, whereas the cross-correlation between sequences is low in all points. Note that the time needed to detect a codeword limits the maximum object speed required for a valid crossing.
IV. TRANSMITTING ANTENNA DESIGN
The key component in the transmitter design is the narrow-beam antenna. In order to satisfy both beam width and side-lobe specifications, a slotted waveguide array [Reference Roselli, Alimenti, Stopponi and Porzi6] has been preferred to easier to use planar antennas because of its lower design uncertainties especially in terms of side-lobe prediction and control. Beam width, gain, and peak-to-side-lobes ratio are parameters that can be controlled by changing the number of elements and the shaping of excitation. Since the radiation pattern is broadside, all the radiators will be driven in phase. The array is of resonant kind, this meaning that the longitudinal distance between slots is half a wavelength at the frequency of operation.
The array is designed according to the method suggested in [Reference Elliot10]. The details of such a design are illustrated in [Reference Alimenti, Mezzanotte, Palazzari, Placentino, Scarponi and Roselli7] and will not be repeated here. The equivalent circuit of the radiating slot is obtained throughout extensive numerical simulation. To this purpose an in-house electromagnetic FDTD code has been exploited. The dimensions of the radiating slot are reported in Table 1, whereas its geometry is modelled as a rectangle with rounded corners.
Table 1. Single radiating cell parameters.
The fabricated linear array is shown in Fig. 2. This structure is composed by 17 slots with optimal Chebyshev shaping. The target peak-to-side-lobe ratio is 25 dB but, as suggested in [Reference Elliot10], such a value is increased by 10 dB in determining the array weights. The measured radiation pattern and reflection coefficient are reported in Fig. 3, whereas the antenna performances are summarized in Table 2. The measured gain is around 15 dB. Since the transmitted power is −13 dBm, it can be concluded that the expected EIRP is about 2 dBm.
Fig. 2. Slotted waveguide array used as transmitting antenna. The required spatial accuracy and side-lobe level is obtained with a 17 slots design.
Fig. 3. Measured radiation pattern (a) and reflection coefficient (b) of the transmitting antenna. The estimated 3 dB beam width is 6°, whereas the side-lobe level is around −25 dB. The operative bandwidth is about 2.5% with an input matching better than −16 dB.
Table 2. Transmitting antenna measured performance.
In order to evaluate the TX antenna performances during production, a low-cost near-field set-up is built. The experimental version of this test bench is shown in Fig. 4. The antenna is placed on a metallic base and connected to the port n.1 of a Vector Network Analyzer (VNA). A very simple near-field probe is obtained by means of a standard SMA-to-microstrip adapter with elongated central conductor. Such a probe is fastened to a planar support, that is able to move on the metallic base, and is connected to the port n.2 of the VNA.
Fig. 4. Photo of the near-field scanning experiment set-up.
The transmission parameter S 21 is then measured while moving the probe along the guide and in front of each slot. The obtained S 21 values are then normalized to their maximum measured at slot n.9 (array center). Figure 5 shows the comparison between the above measurements and those expected from the theoretical Chebyshev weights. The agreement between experimental and theoretical data is quite good even if a slight difference, straddling the theoretical curve, is observed in the considered prototype. This behavior can be ascribed to an error of the slots offset with respect to the antenna central axis (about 6.5%). The error is confirmed by mechanical measurements of the same prototype and causes the obtained weights to oscillate beyond and below the theoretical curve.
Fig. 5. Measured normalized near-field amplitude for each slot compared with the Chebyshev weights.
V. RECEIVING ANTENNA DESIGN
The receiving antenna on the moving target does not need to show high gain, but must be robust, reliable, and light weight. Such a device affects the receiver size and thus its design is geometrically constrained. In automotive applications, not only in F1, for example, one of the few locations for the RFSID tag to be placed is within the envelopes of the external rear-view mirrors.
The adopted solution is a planar antenna array composed by four radiating elements which operates in phase with uniformly distributed currents. In order to reduce the complexity of the exciting network, an input impedance of 100 Ω has been considered. In this way the three power dividers can be realized by simple T-junctions, one of them including quarter-wave impedance transformers.
The radiating cell is formed by a square patch, printed on a low dielectric constant substrate. The exciting point and the dimension of the square have been chosen in order to have an input impedance of 100 Ω at the central frequency. The patch, and in particular the position of the exciting point, has been optimized by means of an electromagnetic simulation based on the FDTD method previously cited. The single-patch and the four-patch array are depicted in Figs 6(a) and 6(b), respectively, while the patch dimensions are summarized in Table 3. The spacing between adjacent elements is assumed to be 0.5λ 0 ≈ 10 mm in order to avoid grating lobes. The resulting directivity is about 12 dB.
Fig. 6. Elementary patch (a) and receiving antenna layout (b). The patch array is realized on a RO-4003 substrate with thickness equal to 0.8 mm and relative permittivity ɛ r = 3.38.
Table 3. Geometrical parameters.
The scattering parameter measurements are reported in Fig. 7, showing an input return loss better than −20 dB in a bandwidth of about 6%. The comparison between the “naked” antenna and that covered by a radome (rear-mirror envelope) are also illustrated in this figure. The measured radiation pattern is shown in Fig. 8, where the received power (normalized to the maximum value) is plotted versus angle. An half-power beam width of about 25° and a peak-to-side-lobes ratio below −20 dB are achieved.
Fig. 7. Reflection coefficient measurement of the receiving antenna with or without radome (3 mm far away the antenna surface).
Fig. 8. Radiation pattern of the receiving antenna measured at 15.4 GHz.
VI. SYSTEM DESIGN
The above system has been realized to operate in Ku-band and the design is based on a low-cost multi-layer PCB (Printed Circuit Board) technology. Commercial mixer and oscillator chips have been widely used in both transmitters and receiver to obtain a highly miniaturized product. The operating frequency can be digitally controlled by PLL (Phase Locked Loop) circuits.
Two digital sub-systems, one for the transmitter and one for the receiver, have been realized adopting low-power mid-density FPGAs. In the transmitter (see Fig. 9 for the transmitter block diagram), the user can select among different gold sequences by means of a switch. The sequences are cyclically generated and then coding is applied thus obtaining a modulating bit stream. In Fig. 10 the transmitter, placed in a real environment (F1 racing circuit path side), is shown.
Fig. 9. Transmitter block diagram.
Fig. 10. Microwave transmitter for lap triggering applications on the field in a F1 track.
The receiver system, the block scheme showed in Fig. 11, is based on a double conversion architecture and exploits the four-patch antenna described in Section V, integrated with the microwave front-end electronics (see Fig. 12).
Fig. 11. Receiver block diagram.
Fig. 12. Microwave receiver: antenna (a) and front-end electronics (b). (A) Via-through; (B) mixer; (C) local oscillator; (D) digital PLL; (E) reference oscillator; (F) second conversion, IF, and signal recovery circuits.
In the receiver a much complex architecture is needed to accurately detect the exact receiving time of each sequence. First, a digital PLL is implemented to recover the clock and then a decoder is inserted. To detect the arrival time of each sequence a correlation receiver has been exploited. A threshold has been finally adopted to detect the correlation peak, corresponding to the arrival time of the input sequence. Table 4 reports the receiver measured characteristics.
Table 4. Receiver measured performance.
Finally, a post-processing logic can be activated to improve the accuracy of the computation of the crossing through the OH line (see Fig. 1). The time of the first received sequence and the time of the last received sequence are stored in the logic and averaged to evaluate the midpoint, then a countdown clock is initialized to give a trigger pulse to the target with a fixed delay after the ideal center-line of the antenna beam.
Due to the adopted architecture and communication scheme, the digital sub-system is able to recognize a transmitted sequence with an absolute accuracy of 2.5 µs. A minimum detection time of 150 µs has been achieved, this corresponding to the duration of at least two sequences.
In the system considered for this particular application the on board unit does not transmit any information back to the fixed ground apparatus (that in conventional RFID systems is seen as the “reader”), but it receives the signal from the ground transmitter and starts recording data on board of the moving object, the first data being the triggering pulse. It is clear that such a system can be easily redesigned and generalized in use to include a back transmitting block on the mobile unit (thus becoming a real transponder), as a classical RFID system would require.
An in-field characterization has been carried out to measure the range of the system. Figure 13 reports a simplified scheme of the measurement set-up. The transmitter is fixed at H = 45 cm above the ground. The receiver is placed on a moving target at the same height (realistic for the reference application). The transmitter continuously sends a set of known data, the receiver compares the received packets with those expected and records the percentage of errors. Once the target is within the covered transmitting area, it starts recording data on board at a constant data rate. If an average transmitted power of −25 dBm is used, no errors are detected up to around 68 m of distance from the transmitter. For −15 dBm of transmitted power, the system range can be improved up to 100 m.
Fig. 13. Schematic view of the range measurement set-up.
At a certain point around 21 m, however, there is a loss of connection. This is due to the multi-path destructive recombination of the transmitted signals along the direct line of site and along the first-order reflection from the soil. The target is not in tracking mode for around 2 m, after that the tracking is recovered. This unwanted effect can easily be mitigated, in the real environment, by paying attention to the positioning of the ground unit.
The system described has been used for years by the top F1 racing teams for lap triggering purpose, giving practical demonstration of the robustness in extreme environmental operating conditions. The receiver was mounted on board of vehicles, whereas the transmitter was placed along the path and no crossing of very fast moving objects (far beyond the wall of 300 km/h) was missing. Although very significant and fascinating, the use described is only one of the several that this system can perform in automotive applications, where, instead of a triggering pulse, a different information can be exchanged between the tag and the reader (i.e. traffic news, alarms, congestion charges, car status, etc.). The system in fact is not flexible only in terms of applicability, but also in terms of its geometrical and mechanical characteristics and potential developments. An example is shown in Fig. 14: in some cases the mechanical constraints of the final user required the tag to be located in very tight sub-environments (such as for instance in the rear-view mirror). In order to reduce the area occupied by the RX antenna and consequently to shrink the size of the whole system, thus allowing for different mounting possibilities, the circuit has been redesigned in a 90° folded version. The figure shows the final industrialized prototype of the folded version, including RF cable and signaling connections between the boards.
Fig. 14. The folded version of the RFSID tag receiver.
VII. CONCLUSION
This paper described a RFSID-based system for road-to-car communication. The system has been validated by several field tests and thousands of proves in a real harsh environment, demonstrating a full functionality, flexibility, and robustness. The maximum measured range is over 90 m while overall accuracy better than 1 ms has been experienced. It has been extensively used as “lap trigger” in F1 competitions with a fault rate close to zero. From these results, it can be concluded that the proposed Ku-band RFSID-based system is a robust solution for future applications, where robustness, reliability, and precision in location-based communication are main constraints.
ACKNOWLEDGEMENTS
This paper was partially supported by the COST Action IC0803 “RF/Microwave Communication Subsystems for Emerging Wireless Technologies (RFCSET).”
Luca Roselli received the Laurea degree in electronic engineering from the University of Florence, Florence, Italy, in 1988. From 1988 to 1991 he worked at the University of Florence on SAW devices. In November 1991, he joined the Institute of Electronics at the University of Perugia, Italy, as a research assistant. Since 1994 until 1998 he had been teaching “Electronic Devices” at the same University. Since 1998 he has been teaching “Microwave electronics.” He is in the reviewer list of the IEEE Microwave and Wireless Component Letters and in that of the IEEE transactions on Microwave Theory and Techniques. Since 1998 he is a member of the Technical Program Committee (TPC) of the International Microwave Symposium. In June 2000 he was appointed as “Associated Professor” in applied electronics at the University of Perugia; since that time he has been coordinating the research activity of the High Frequency Electronics (HFE) Lab. In the same year he founded the WiS (Wireless Solutions) S.r.l., a spin off company operating in the field of microwave electronic systems with which he is currently cooperating as a consultant. In 2003 he was a member of the spin off committee of the University of Perugia, in 2005 he founded a second spin off company: DiES (Digital Electronic Solutions) S.r.l. In 2007 he was the chairman of the VII Computational Electromagnetic in Time Domain workshop, in 2008 he was a member of the Steering Committee of the International Microwave Symposium (IMS) held in Atlanta; he was the guest editor of the special issue of the International Journal of Numerical Modeling on the VII CEM-TD. In the same year he was nominated as representative for Italy in the Management Committee of COST Action IC0803, RF/Microwave Communication Subsystems for Emerging Wireless Technologies (RFCSET) where he is currently operating as Work Group leader. Again in 2008 he was the Co-PI (CO-Principal Investigator) of the Project ADAHELI (funded by the Italian Space Agency) for space-based solar flares.
Valeria Palazzari received the Laurea degree in electronic engineering and the Ph.D. degree in information and electronic engineering from the University of Perugia, Italy, in 2000 and 2003, respectively, under the advice of Prof. Luca Roselli. Since 2000 she joined the HFE (high frequency electronics) research group at the Department of Electronic and Information Engineering (DIEI) at the same university working as young research assistant. In summer 2003 she joined the Georgia Institute of Technology (GA) ATHENA research group within a Summer International Fellowship program sponsored by PRC under the advice of Prof. Manos M. Tentzeris. Her work was mainly focused on the design, realization, and test of WLAN asymmetrical dual-band filters implemented on LCP. She is currently working as a post-doc assistant at DIEI, and teaching telecommunication electronics within the information and telecommunication engineering course of the Univ. of Perugia in Orvieto (TR), Italy. She is also a R&D engineer at WiS S.r.l. (PG), Italy. Her research goes from the modeling, design, and realization of microwave integrated circuits in Si/SiGe BiCMOS technology, to packaging technologies for RF/wireless systems. In 2008 she was leader of the “MIOS– scientifical requirements definition” work package within the ADAHELI project, a small space mission funded by the Italian Space Agency (ASI) and devoted to the observation of the Sun.
Federico Alimenti received the Laurea degree (summa cum laude) and the Ph.D. degree from the University of Perugia, Italy, in 1993 and 1997 respectively, both in electronic engineering. In 1993 he held a scholarship from Daimler Benz Aerospace, Ulm, Germany. In 1996 he was awarded as young scientist from URSI (Union Radio-Scientifique Internationale) and he was visiting scientist at the Lehrstuhl für Hochfrequenztechnik of the Technical University of Munich, Munich, Germany. Since 2001 Federico Alimenti has been with the Department of Electronic and Information Engineering of the University of Perugia as an assistant professor, teaching the classes of telecommunication electronics. Since 2004 he belongs to the TARGET (Top Amplifier Research Group in a European Team) Networks of Excellence and in this frame he is involved in the design of Microwave Power Amplifiers in Si/SiGe BiCMOS Technology. Since 2002 he is consultant of industry. Since 2006 he is the researchers' deputy at the Academic Senate of the University of Perugia. Since 2007 he is reviewer of the European Microwave Integrated Circuits Conference – EuMIC. In 2007 he has been steering committee member for the “7th Workshop on Computational Electromagnetics in Time Domain CEM-TD,” a conference sponsored by IEEE. He is IEEE member. In 2008 he was leader of the “MIOS Preliminary Design” Work Package within the ADAHELI project, a small space mission founded by the Italian Space Agency (ASI) and devoted to the observation of the Sun. His research interests concern the modeling, design, and realization of microwave integrated circuits in CMOS and Si/SiGe BiCMOS technologies. In addition, he has an experience on the design and testing of microwave radiometers. He is IEEE member, serving as reviewer of the following journals: IEEE Microwave and Wireless Component Letters; IEEE Transaction on Microwave Theory and Techniques; IEEE Transaction on Advanced Packaging; and IEEE Transaction on Circuit and System I.
Paolo Mezzanotte received the Ph.D. degree from the University of Perugia, Italy, in 1997. Since 1992, he was involved with FDTD analysis of microwave structures in cooperation with the Department of Electronic and Information Engineering (DIEI), University of Perugia. In 1999, he was appointed Research associate. Since January 2007, he is an Associate Professor with the same University, teaching the classes of microwave and radiofrequencies. His research activities concern numerical methods and CAD techniques for passive microwave structures and the analysis and design of microwave and millimeter-wave circuits. More recently, his research interests were mainly focused on the study of advanced technologies such as LTCC and RF-MEMS. Paolo Mezzanotte was the co-chairman of the seventh edition of the international workshop “Computational Electromagnetics in Time-Domain” (CEM-TD 2007).
Matteo Comez received the Laurea degree in electronic engineering from the University of Perugia, Italy, in 1999. Presently, he is a program manager at WiS S.r.l., Perugia, Italy.
Nicola Porzi received the Laurea degree in electronic engineering from the University of Perugia, Italy, in 1999. Presently, he is a production chief engineer at WiS S.r.l., Perugia, Italy.
