A Comparison of ZigBee/ vs. Ultra-Wideband Transmission Techniques for Wireless Fire and Intrusion Detectors

Transcription

1 Thomas Kaiser Universität Duisburg-Essen, Duisburg, Germany Jianliang Shi Fraunhofer Institute for Microelectronic Circuits and Systems, Duisburg, Germany Nima Tavangaran, Ingolf Willms Universität Duisburg-Essen, Duisburg, Germany A Comparison of ZigBee/ vs. Ultra-Wideband Transmission Techniques for Wireless Fire and Intrusion Detectors Abstract Recently, wireless sensor networking has been becoming a disruptive technology in such fields as industrial control and monitoring, building automation and security, automotive sensing. The motivation for use of wireless technology includes the reduction of installation cost, the avoidance of the last meter connectivity problem and intelligent maintenance. Many types of wireless sensors, for example fire and intrusion detectors, are based on proprietary solutions, which are uncomfortable for the customer because of lack of product interoperability and vendor dependence. On the other hand, the lack of standardized technologies that can address the requirements both at the application level and from the communication point of view has become the hurdle of the widespread of wireless sensor systems. The above constraints suggest the necessity of an open global standard to enable the cost-effective, reliable wireless monitoring and control products. The recently finished IEEE standard and the corresponding undergoing Zigbee standard represent the achievements of the efforts addressing these kinds of requirements. Zigbee/ operates at 868/915MHz, or 2.4GHz and belongs to classical narrowband transmission systems. An entire opposite transmission approach is based on extremely low power signals with ultra-wide bandwidth (UWB). UWB belongs to the few wireless key emerging technologies and is under intense discussion for various areas of applications, e.g. highest data rates in the order of several hundred MBit/s for indoor communications, ranging and positioning with accuracy in the sub-centimeter range, and recently also for low power and low data rate applications.

2 Aim of this contribution is to compare narrowband vs. ultrawideband transmission with special emphasis on wireless fire and intrusion detector networks. Introduction The Zigbee/ standards, with such features as ultra low complexity, ultra low power and low bit rate, are ideally suited for wireless fire and intrusion detectors. Numerous wireless fire and intrusion detectors are based on proprietary solutions, which are uncomfortable for the customer because of lack of product interoperability and vendor dependence. From this point of view ZigBee seems to be the right choice in order to serve future market demands. On the other hand, in early 2004 the Institute of Electrical and Electronics Engineers (IEEE) Working Group voted to form a new study group to investigate amendment to for an alternative physical layer with additional features like precision ranging/location capability, high aggregate throughput, scalable data rates, and extremely low power consumption. All the features can be principally achieved with Ultra-Wideband Technology. The aim of this contribution is to shed some light on the pros and cons of Zigbee vs. UWB with emphasis on their application for wireless fire and intrusion detectors. For example, because ZigBee is a conventional narrowband technique operating with a carrier frequency of either 2.4 GHz (Global), 915 MHz (North American) or 868 MHz (Europe) it suffers from other narrowband interferences as well as the known fading phenomenon. In contrast, UWB may not only mitigate narrowband interferers simply by notch filtering or other advanced signal processing techniques, but also mitigates fading by non-destructive impulsive transmission. For both approaches, the very low duty cycle guarantees low power consumption and co-existence, but the lack of severe fading in UWB transmission may result in a reduced path loss and consequently in a larger coverage. However, according to current American regularization, UWB systems are allowed to operate mainly in the frequency range between 3.1 GHz and 10.6 GHz with a maximum power spectral density of dbm/mhz, resulting in a maximum transmit power of less than 0dBm. Compared to Zigbee, this higher frequency range causes additional attenuation by walls and other obstacles and therefore limits the coverage gain. In turn, due to long spreading sequences, UWB devices can principally be designed to work below the noise floor so that jamming becomes extremely difficult.

3 Such a property is beneficial especially for intrusion detectors. Note also the pulse duration in the sub-nanosecond range providing a fine time resolution and enabling precise ranging with centimeter resolution. All these facts open the question whether UWB represents a valuable alternative to standardized narrowband transmission systems. The following section summarizes the properties of UWB whereas the next section deals about Zigbee/ Then, the individual link budgets are calculated and compared for different carrier frequencies, typical data rates in wireless fire and intrusion detectors, and bandwidths in order to allow a quantitative comparison. A conclusion and an outlook finalize this contribution. The Ultra-Wideband Approach Ultra-wideband systems are considered as systems that are operating with a large absolute bandwidth of more than 500 MHz. Using a large bandwidth in UWB systems provide us with several possibilities and advantages compared to other narrowband systems. First, the spectral density of the transmit signal can be made rather small and consequently the probability of interference to narrowband systems is greatly reduced. Immunity to narrowband interference, mitigating fading, high data rate and a better propagation in different environments are among other advantages of UWB systems [1]. In February 2002, the American Federal Communication Commissions (FCC) issued its first report and order on UWB technology, thereby providing regulations to support deployment of UWB radio systems. After that, a great deal of efforts and money has been invested in finding solutions for implementation of this technology in practice. Modulation The most popular current modulation schemes are Pulse-Position Modulation (PPM), Pulse-Amplitude Modulation (PAM) and Orthogonal Frequency Division Multiplex (OFDM), whereas the extensions Direct Sequence (DS) and Time Hopping (TH) enable multiple accesses to the common transmission medium. PPM encodes the data through different positions in time and PAM encodes the data through signal amplitudes. TH assigns to multiple different codes, where the codes in turn represent a particular pattern in a series of pulses per transmitted bit, i.e. the transmitted signal in time hopping is represented by

4 s tr i= ( t) = a x ( t it c T ), i N f tr f i c where x tr (t) is the transmitted pulse, T f the frame duration, T c pulse duration or chip interval, N f the number of pulses representing a single bit, c i the periodical code word for each user and finally a i N f = ± 1 represents the binary information hidden in the pulse train and is the floor operator giving the largest integer smaller than or equal to its argument. The transmitted pulse is often simplified as a Gaussian mono-pulse [ ] [( t ) τ ( t T ) τ e T c ] 2 xtr ( t) = k c, where k is a normalizing factor and τ is the monocycle s duration. One way to satisfy the FCC-mask requirements is taking derivatives of x tr (t) and adequately adjusting the center frequency. In UWB-OFDM, the regulated frequency band is divided into several sub-bands with individual bandwidth of 500 MHz and sub-dividing each sub-band into 128 tones, each with a bandwidth of approx. 4 MHz. UWB-OFDM permits adaptive selection of sub-bands in order to fulfill regional regulations requirements or for reasons of co-existence. The Zigbee/ Standard The purpose of IEEE is to provide a standard for ultra low complexity, ultra low cost, ultra low power consumption and low data rate wireless connectivity devices. It defines a standard for a Low Rate Wireless Personal Area Network (LR_WPAN) working in the industrial, scientific, and medical (ISM) bands. The scope of the standard is to define the physical layer (PHY) and medium access control (MAC) sublayer specifications for low data rate wireless connectivity with fixed, portable, and moving devices with no battery or very limited battery consumption requirements typically operating in the personal operating space (POS) of 10m [2].ZigBee is an alliance of semiconductor manufacturers, technology providers, and OEM's dedicated to providing the upper layers of the protocol stack (from network to application, including application profiles) atop the IEEE PHY and MAC layers [4]. The ZigBee protocol together with IEEE standard provides an open standard for low-power wireless networking of monitoring and control devices.

5 Highlights of IEEE standard As noted before, the goal of IEEE standard is to provide 4- (ultra) low reliable wireless connectivity. This section presents some unique features of the standard that enable the performance vision to be met. First of all, IEEE standard works in ISM band and its service involves little or no infrastructure. This feature allows power-efficient, inexpensive solutions to be implemented for a wide range of devices. Other notable features of the standard will be described as follows. Modulation and related features IEEE provides two PHY options: one for 868/915 MHz bands, and the other for 2.4 GHz band. The 868 MHz band is available in most European countries; the 915 MHz band is available in North America, and the 2.4GHz band is available in most countries worldwide. Table 1 summarizes the modulation formats of different frequencies. As is shown in Table 1, both PHYs use direct sequence spread spectrum (DSSS) methods, which results in low-cost IC implementation via largely digital circuits. Data modulation schemes are simple raised-cosine-shaped BPSK for 868/915 MHz PHY, and half-sine-shaped O-QPSK in the 2.4GHz PHY. Each of these modulation schemes maintains a peak-to-average carrier power ratio of one, which minimizes both power consumption and implementation complexity [3]. PHY (MHz) 868/ 915 Freq. band (MHz) Spreading parameters Chip Modulation rate (kchip/s) Bit rate (kb/s) Data parameters Symbol Symbols rate (symbols /s) 300 BPSK Binary BPSK Binary O-QPSK ary orthogonal Amount of channel 16 Table 1 Frequency bands and data rates Further, the 868 and 915MHz bands are considered close enough in frequency that similar, if not identical, hardware can be used for both, lowing the implementation costs

6 [5]; the 2.4GHz PHY supports 16 channels in an 83.5MHz bandwidth, which means a channel spacing of 5MHz, and this greatly eases the transmit and receiver filter requirements. Coexistence features IEEE standard provides several mechanisms that enhance coexistence with other wireless devices operating in the same frequency band, among them, channel selection/ access and its supporting functions, low transmit power, low duty cycle, modulation and channel alignment are notable mechanisms. The ability to detect channel occupancy and perform dynamic channel selection is an important mechanism for coexistence. The IEEE MAC layer includes a scan function, which, when performing dynamic channel selection, either at network initialization or in response to an outage, will scan a set of predefined channels. PHY layers, on the other hand, contain several lower-level functions, such as receiver energy detection (ED), link quality indication (LQI), and channel switching, which make the channel relocation possible. The mechanisms provided by both PHY layer and MAC layer are intended for use as part of a channel selection algorithm at the network layer. IEEE PHYs provide the capability to perform clear channel assessment (CCA) in its CSMA-CA mechanism. The CCA is based on: ED over a threshold, detection of IEEE packet, or both. Use of the ED option improves coexistence by allowing transmission backoff if the channel is occupied by any device using any protocol. Among other coexistence features, low duty cycle (under 1%) obviously will make the devices less likely to cause interference to other standards; the operating power of IEEE devices are expected to be between 3 and 10dBm, with 0 dbm being typical. This is much less than other kind of devices and therefore will cause less interference. The use of DSSS not only improves the performance of the receiver by reducing the impact of interference from other kind of devices, but also causes less interference to others. Low power consumption features In many applications that use IEEE standard, the devices will be battery powered where their replacement or recharging in relatively short intervals is

7 impractical; therefore the power consumption is of significant concern. This standard was developed with the limited power supply availability in mind. The protocol has been developed to favor battery-powered devices. The PHY layer of the IEEE standard provides numerous features which enable the low active power consumption, for example, DSSS, constant envelope modulation, no duplex operation, etc [6]. Although the specification has been designed to minimize the active power consumption of compliant devices, it is impossible to reach the goal of at least a few months of operation with a single battery by reducing the active power alone. Most first-generation IEEE chips will consume ma from a 2 V supply (40 60 mw) while active, which indicates that low power consumption is achieved largely by low-duty-cycle operation. To support low duty cycles, the IEEE MAC layer enables beacon packet being as short as 544us in 2.4GHz band, while the superframe period may be extended from 15.36ms to over four minutes. This results in a duty cycle that may be set from 2.3% to 2.16 ppm[3]. Furthermore, non-beacon mode enables a device to remain in a standby mode indefinitely unless otherwise to satisfy the regulatory requirements. Another MAC layer power saving feature includes battery life extension (BLE) mode, in which the CSMA-CA backoff exponent is limited to range 0-2. This greatly reduces receiver duty cycle in low offered traffic applications. A Link Budget Based Quantitative Comparison A quantitative comparison between UWB and ZigBee link budgets gives us a better insight on the applicability of each system for wireless low power and low data rate systems. First of all we consider the UWB frequency range 3.1 GHz GHz and compare the link-budgets for both systems. Then, we alternatively investigate the UWB frequency range 0.5 GHz - 1 GHz while being aware of potential interferers like GSM or TV signals. Table 2 shows a quantitative comparison of UWB and ZigBee link budgets with a tentative throughput of 2400bps where the whole bandwidth between 3.1 GHZ and 10.6 GHz for UWB is taken into account. Note that the pathloss coefficient is more or less dependent on the bandwidth. Here, we assume an average pathloss coefficient of 2.5 for

8 UWB and 3.5 for a narrowband signal, which is in well agreement with published measurement results [7]. The total pathloss can be modelled by: 4πd 0 d PL = 20log α log10, λ d 0 where d = 0 1m, λ is the wavelength, α is the pathloss coefficient and d is the distance between transmitter and receiver which is fixed to 10m in Table 2. The received power Pr = Pt + Gt + Gr PL depends on the transmitted power Pt, the transmit Gt and receive Gr antenna gains and the path loss. Thermal noise is mainly determined by the low noise amplifier at the receiver N 10 = 174 dbm + 10log Rb, where Rb represents the throughput. The total average noise power per bit is given by Pn = N + Nf, where Nf is the receiver s noise figure. The processing gain ( PG ) is defined as PG = 10log( spreading factor ), where the spreading factor is the number of pulses used to decode a single bit. Finally, the targeted link margin is calculated by M = Pr Pn S I + PG, where S and I are minimum E b N0 implementation loss, respectively. E b N0 is the measure of signal to noise ratio for a digital communication system. It is simply defined as the ratio of Energy per Bit ( E ) to the Spectral Noise Density ( N 0 ). Exploiting the full UWB bandwidth with a high center frequency of 6.85 GHz causes a significant pathloss compared to ZigBee. According to Table 2, the total pathloss for UWB at 10m is around 74.2 db, while for ZigBee with 868 MHz center frequency is only 66.2 db. i and b Consequently, decreasing the center frequency and keeping the bandwidth at its minimum of 500 MHz, i.e. considering the range 3.1GHz-3.6 GHz, might present an alternative approach. The corresponding link budget is shown in Table 3 and the pathloss (PL) as well as its distance to the processing gain (PG) in Figure 1. Only 6.3 db improvement in total pathloss can be gained which is still far away from ZigBee s counterpart in terms of pathloss. Note that the quantity PG-PL is a good measure for the signal power available at the detector for data reconstruction and takes into account possible signal processing at the receiver and is therefore rather meaningful because of direct proportionality to the bit error rate.

9 Hence, if no signal processing is present at the receiver (e.g. due to cost constraints), Zigbee/IEEE outperforms UWB up to a distance of approximately 10m. Otherwise, if the receiver does include signal processing, UWB shows better performance for distances greater than 3m. UWB ZigBee Geometric center frequency (Fc) MHz Throughput (Rb) bps 3-dB bandwidth MHz Alpha (Pathloss exponent) Average transmit power (Pt) dbm Transmitted power for 1 MHz dbm Path loss at 1m db Total Path loss at 10 m db Rx power at 10m (Pr) dbm Average noise power per bit (N) dbm Rx noise figure (Nf) 7.0 [9] 7.0 db Total Ave. noise power per bit (Pn) dbm Minimum E b /N o (S) 3.6 [9] 8.7 db Implementation loss (I) 3.0 [9] 3.0 db Processing Gain (PG) db Link Margin at 10m db Table 2 UWB (3.1GHz-10.6 GHz) and ZigBee link budget comparison This preliminary conclusion motivates us to consider an even lower frequency-band like 0.5 GHz-1 GHz. Before doing any calculation we first take a closer look at the radio spectrum to see which services are being offered in this frequency range. The main interfering services are TV broadcasting signals which occupy in the United States a huge bandwidth from MHz. Cellular phones, biomedical telemetry devices and navy radars are other examples of interfering services [8]. Coming back to our comparison, Table 4 shows the link-budget for both systems where UWB is now limited to the 0.5GHz-1 GHz range. The total pathloss for UWB is 54.9dB and for ZigBee 66.2dB which finally shows a distinct improvement for UWB. Figure 2 gives a better feeling how fast the pathloss for ZigBee grows comparing to UWB and finally results in a difference of 11.3dB at a distance of 10m. Most interestingly, we can

10 see this result in the received power, where for both UWB and ZigBee, the received power is exactly the same (-69.2 dbm) while ZigBee suffers from a much larger transmit power (-14.3 dbm for UWB and -3 dbm for ZigBee). UWB ZigBee Geometric center frequency (Fc) MHz Throughput (Rb) bps 3-dB bandwidth MHz Alpha (Pathloss exponent) Average transmit power (Pt) dbm Transmitted power for 1 MHz dbm Path loss at 1m db Total Path loss at 10m db Rx power at 10m (Pr) dbm Average noise power per bit (N) dbm Rx noise figure (Nf) db Total Ave. noise power per bit (Pn) dbm Minimum E b /N o (S) db Implementation loss (I) db Processing Gain (PG) db Link Margin at 10m db Table 3 UWB (3.1GHz-3.6 GHz) and ZigBee link budget comparison UWB PL ZigBee PL UWB PG-PL ZigBee PG-PL Distance (m) Fig. 1 Pathloss and its difference to processing gain for UWB and Zigbee/ IEEE as a function of distance and frequency range 3.1GHz-3.6 GHz

11 UWB ZigBee Geometric center frequency (Fc) MHz Throughput (Rb) bps 3-dB bandwidth MHz Alpha (Pathloss exponent) Average transmit power (Pt) dbm Transmitted power for 1 MHz dbm Path loss at 1m db Total Path loss at 10m db Rx power at 10m (Pr) dbm Average noise power per bit (N) dbm Rx noise figure (Nf) db Total Ave. noise power per bit (Pn) dbm Minimum E b /N o (S) db Implementation loss (I) db Processing Gain (PG) db Link Margin at 10m db Table 4 UWB (0.5-1 GHz) and ZigBee link budget comparison UWB PL ZigBee PL UWB PG-PL ZigBee PG-PL Distance (m) Fig. 2 Pathloss and its difference to processing gain for UWB and Zigbee/ IEEE as a function of distance and the frequency range GHz

12 Observe that now UWB outperforms Zigbee/IEEE for distances larger than 10cm (1m) if signal processing at the receiver is (not) carried out, respectively. Conclusions and Outlook In this paper, we conduct a comparison study of Zigbee/IEEE technology vs. ultra-wideband technology with application to wireless sensor networking, particularly wireless fire and intrusion detection. The details of both technologies are introduced and their general characteristics extracted and compared. In particular, a quantitative comparison is carried out by studying the individual link budget of Zigbee/IEEE and that of UWB at different carrier frequencies and bandwidth s. Our results show that decreasing the frequency range down to 0.5GHz-1.0GHz while keeping the bandwidth at 500MHz, pathloss of UWB and Zigbee/IEEE are in the same order. However, the more relevant measure PG-PL favors UWB for distances larger than a few centimeters. In turn, for the conventional frequency region 3.1GHz-3.6GHz, Zigbee/IEEE outperforms UWB for distances smaller than 3m (10m), if signal processing at the receiver is (not) taken into account, respectively. Both Zigbee/ and UWB technologies share some attractive features such as ultra low power consumption, low cost and coexistence. The UWB technology holds some further features like multi-path mitigation, precision location and immunity to interception. From this point of view, we believe that both technologies will be the wave of the future of wireless sensor networking, where UWB is advantageous for wireless fire and intrusion detection if the location of the detector becomes relevant or if the receiver can be equipped with signal processing. For example, jamming in UWB is already rather difficult because of the low spectral density, but the UWB inherent ranging facility further complicates jamming. This might be beneficial in particular for intrusion detectors. Otherwise, for extremely low cost fixed sensor networks with insignificant sensor positions and small sensor spacings, Zigbee/IEEE seems to be the favorite technique. However, note that UWB allows rather cheap transmitters because mixer and oscillator are not required. Moreover, a remarkably battery saving can be expected from UWB based sensor networks, provided that the mentioned interferers can be

13 significantly suppressed. Other related criteria, like antenna size restrictions, need to be investigated further. In conclusion, if striving for a product, all hardware related criteria and not only the link budget have to be taken into account for selection of the most adequate technological approach. References [1] A. F. Molisch, Y. G. Li, Y. Nakache, P. Orlik, M. Miyake, Y. Wu, S. Gezici, H. Sheng, S. Y. Kung, H. Kobayashi, H. Vincent Poor, A. Haimovich, and J. Zhang, A low-cost time-hopping impulse radio system for high data rate transmission, to appear in EURASIP Special Issue UWB State of the Art, October 2004 [2] IEEE Std , IEEE Standard for Information technology Telecommunications and information exchange between systems Local and metropolitan area networks Specific requirements Part 15.4: Wireless Medium Access Control (MAC) and Physical Layer (PHY), Specifications for Low- Rate Wireless Personal Area Networks (LR-WPANs). [3] J. A. Gutierrez, E. H. Callaway, R. L. Barrett, Low-rate Wireless Personal Area Networks: Enabling Wireless Sensors with IEEE Standards Informaiton Network, IEEE Press, [4] Homepage of ZigBee Alliance, [5] E. Callaway, P. Gorday, L. Hester, J.A. Gutierrez, M. Neave, B. Heile, V. Bahl, "Home networking with IEEE : A developing standard for low-rate wireless personal area networks," IEEE Communication Magazine, vol. 40, no. 8, pp , August [6] E. Callaway, Low power consumption features of the IEEE /Zigbee LR- WPAN standard, Mini-tutorial, ACM Sensys 03, Los Angeles, CA, USA, November 5-7, [7] S. Ghassemzadeh, V. Tarokh, The Ultra-Wideband Indoor Pathloss Model, IEEE P Working Group for Wireless Personal Area Networks (WPANs), July, 8, 2002 [8] The Federal Communications Commission s table of frequency allocations, 47 C.F.R , Revised on April 13, 2004 [9] G. R. Aiello, G. D. Rogerson, Ultra-Wideband Wireless Systems, IEEE microwave magazine, June 2003