Wireless Communications In Tunnels For Urban Search And .
View metadata, citation and similar papers at core.ac.ukbrought to you byCOREprovided by CiteSeerXWireless Communications in Tunnels for Urban Searchand Rescue Robots *Kate A. Remley, Galen Koepke,Dennis G. Camell,Chriss GrosvenorNIST Electromagnetics Division325 BroadwayBoulder, CO, USA 80305 1 303 497-3652Lt. George HoughRobert T. JohnkNew York City Fire DepartmentNew York City, NY USAInstitute for TelecommunicationsSciences325 BroadwayBoulder, CO, USA @boulder.nist.govABSTRACTWe report on propagation tests carried out in a subterraneantunnel to support improved wireless communications for urbansearch and rescue robots. We describe single-frequency andultrawideband channel-characterization tests that we conducted,as well as tests of telemetry and control of a robot. We utilizepropagation models of both single-frequency path loss andchannel capacity to predict robot performance. These models canalso be used for optimizing wireless communications in tunnels ofvarious sizes, materials, and surface roughness.Categories and Subject DescriptorsDefining and measuring aspects of an intelligent system.Evaluating components within intelligent systemsGeneral TermsMeasurement, Performance,Standardization, ultipath; radiowave propagation; RMS delay spread; robot;urban search and rescue; wireless communication; wirelesssystemINTRODUCTIONResearchers from the Electromagnetics Division of the NationalInstitute of Standards and Technology (NIST) and the FireDepartment of New York (FDNY) conducted tests of radiowavetransmission and detection in tunnels at the Black Diamond MinesRegional Park near Antioch, California on March 19-21, 2007.Our goal was to investigate propagation channel characteristicsthat affect the reliability of wireless telemetry and control ofUrban Search and Rescue (US&R) robots in tunnels and otherweak-signal environments. We describe measurement methodsthat we used to study parameters relevant to robot performance.We also use the measured data to verify models of radiowavepropagation in tunnels. These models can be used to predict robotperformance in tunnels having characteristics different from theones we measured, such as subways and utility tunnels. This worksupports the development of technically sound standards forUS&R robots [1-3].We used both time- and frequency-domain techniques to studyissues such as channel multipath and loss that may impedesuccessful wireless communications in tunnels. We tested bothvideo and control of a robot inside a mine tunnel. We alsoimplemented propagation models of path loss and channelcapacity and compared our measured results to these models. Wesummarize below the data we collected and interpret the keyfindings from the study, which is described in its entirety in [4].Recently, the wireless field has seen a renewed interest instudies of signal propagation in both mine and subway tunnels,following a good deal of study on mine communications in the1970s. A seminal work on mine tunnel propagation by Emslie etal. [5], studied radiowave propagation in small underground coaltunnels (4.3 m wide x 2.1 m high) for frequencies ranging from200 MHz to 4 GHz. Emslie developed a model for propagation intunnels that is used today. Recently, Rak and Pechak [6] appliedEmslie’s work to small cave galleries for speleologicalapplications, confirming Emslie’s findings that once a fewwavelengths separate the transmitter and receiver, the tunnel actsas a waveguide that strongly attenuates signals below thewaveguide’s cutoff frequency. Because the walls of the tunnel arenot perfectly conducting, signals operating above the cut-offfrequency also experience significant loss. In a recent paper,Dudley, Lienard, Mahmoud, and Degauque [7] performed adetailed assessment of operating frequency in a variety of tunnels.They found that as frequency increases, the lossy waveguideeffect decreases.Our measurements, covering a much wider frequency rangethan [7], and implementation of the model of [6] also confirm thelossy waveguide effect in the tunnels we studied. This effect canhave a significant impact on choice of frequency for criticalapplications such as US&R operations, where typicallyinfrastructure such as a repeater network is not available and livesmay be at stake.Another factor in tunnel communications is multipath causedby reflections off the walls, floor, and ceiling of the tunnel. Thiswas clearly seen in the work of Dudley, et al. [7]. Multipath canhave a pronounced effect on successful transmission of wideband*Publication of the U.S. government, not subject to copyright in the U.S.Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not madeor distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, topost on servers or to redistribute to lists, requires prior specific permission and/or a fee.PerMIS’08, August 19–21, 2008, Gaithersburg, MD, USA.Copyright 2008 ACM 978-1-60558-293-1 5.00
data. Multipath interference may affect certain frequencies in awideband signal while simultaneously having little impact onother frequencies. This frequency selectivity can make decodingsignals difficult for the demodulator in a receiver.During the tests, the radio antennas were approximately 0.75 mfrom the floor, a height similar to that of the robot we studied.We studied the severity of multipath in the tunnel environmentby measuring the RMS delay spread, as well as the success orfailure of wideband data transfer by use of a commerciallyavailable robot. We compare our measured results to a model ofchannel capacity based on Shannon’s theory of channel capacity[6]. This theorem provides a basis for predicting the success ofwireless communication in multipath environments.We first describe the measurements we carried out andinterpret the results. We then discuss the models we implementedand how they can be used to optimize radio communications forUS&R applications.GreathousePortalHazel-AtlasPortal1. TEST ENVIRONMENT(a)The Black Diamond Mines are part of an old silica minecomplex that was used early in the 1900s to extract pure silicasand for glass production. As such, the walls of the mine shaftsare rough and consist of sandy material.Two tunnels were studied, the Hazel-Atlas North (here calledthe “Hazel-Atlas” tunnel) and Hazel-Atlas South (here called the“Greathouse” tunnel). The tunnels are located beneath a mountainand are joined together deep inside, as shown in Figure 1(a). Thedimensions of the Hazel-Atlas tunnel varied from approximately1.9 m (6’, 3”) x 1.9 m to as much as 2.6 m (8’, 5”) x 2.4 m (8’,0”). The dimensions of the Greathouse tunnel were somewhatbigger, up to approximately 3 m square in places. The HazelAtlas tunnel contained railroad tracks spaced 61 cm (24”) apart.Both tunnels consisted of a straight section followed by a 90degree turn around a corner, as shown in Figure 1(a). Below wereport on results for the Hazel-Atlas mine tunnel, shown in moredetail in Fig. 1(b). The complete set of data on both tunnels can befound in [4].Figure 2 shows photographs of the tunnel, including: 2(a) theportal (entrance) of the Hazel-Atlas mine; 2(b) approaching theright-angle turn shown in Figure 1(b); and 2(c) past the turn. Thephotos show the rough, uneven walls in the tunnels, some withwooden shoring, and the railroad tracks.2. MEASUREMENTS2.1 Narrowband Received PowerWe measured the power received outside the tunnel from atransmitter placed at various locations inside the tunnel. Wecollected single-frequency (unmodulated, carrier only) receivedpower data at frequencies near public-safety bands(approximately 50 MHz, 150 MHz, 225 MHz, 450 MHz).Gathering information at these frequencies helps to provide achoice of optimal frequency for the US&R community for thisenvironment, both for robot communications and for other typesof radio communication. These data provide insight into the lossywaveguide effect mentioned in the Introduction.The handheld transmitters we used were radios similar tothose of first responders, except they were placed in ruggedizedcases and were modified to transmit continuously. Each radiotransmitted a signal of approximately 1 W through anomnidirectional “rubber duck” antenna mounted on the case.200180160#3140#2/#4120100#1/#5RX(b)Figure 1: (a) Overview of the Hazel-Atlas mine tunnelcomplex. The network of mines is located deep within amountain. The dark-shaded areas are accessible. (b) Close-upview of the Hazel-Atlas tunnel. The dashed line shows the pathalong which we took measurements, including the right-angleturn at 100 m. The triangles indicate the distance in meters, theovals correspond to locations shown in Fig. 3, and thereceiving equipment is labeled RX.We carried the radio transmitters from the entrance tolocations deep within the tunnels while continuously recordingthe received signal. From the Hazel-Atlas tunnel portal, wemoved the transmitter approximately 100 m down a straighttunnel, then turned a corner and proceeded another 100 m, asshown in Fig. 1(b).The receiving equipment was located just outside the portal.Omnidirectional discone receiving antennas were mounted ontripods, as shown in Fig. 2(a). We used a narrowbandcommunications receiver to convert the received signal to audiofrequencies, where it was digitized by a computer sound card andrecorded on a computer. This instrument, when combined with
NIST-developed post-processing techniques [1, 8], provides ahigh-dynamic-range measurement system that is affordable formost public-safety organizations. Part of the intent of this projectwas to demonstrate a user-friendly system that could be utilizedby US&R organizations to assess their own unique propagationenvironments.Figure 3 shows representative measured received-power dataat frequencies of 50 MHz, 162 MHz, and 448 MHz acquiredwhile the transmitters were carried by foot through the tunnel.The signals were sampled at approximately 48 kHz and the poweraveraged over one-second intervals. The left and right halves ofthe graph show measurements made walking into and out of thetunnel, respectively, and thus mirror each other. The verticaldashed lines on the graph correspond to the entrance (#1, #5), turn(#2, #4), and turn-around point (#3) in the measurement path, asshown in Fig. 1(b).The cut-off frequency for this type of tunnel is difficult todefine since the walls behave as lossy dielectrics rather thanconductors. These conditions are discussed in [9], where theattenuation constant is found to vary as the inverse of frequencysquared [Section 2.7, pp. 80-83]. Hence, we would expect higherattenuation at the lower frequencies but no sharp cut-off. Furthercomplications in this tunnel are the axial conductors (cables,water pipes, rails) that may support a TEM-like mode ofpropagation, the irregular cross-section, and the side chambersand tunnels.For the Hazel-Atlas mine tunnel, we see in Fig. 3(a), strongattenuation of the 50 MHz signal and in Fig. 3(b), the receivedpower of the 162 MHz signal also decreases rapidly as thetransmitter moves into the tunnel. This rapid attenuation is due tothe lossy waveguide effect described in references [4-7]. Thesignal for the 448 MHz carrier frequency (Fig. 3(c)) exhibits lessattenuation and this is where the models of [5] may apply. Signalsmay travel even further at higher frequencies, as discussed in [57]. This frequency dependence may play a significant role indeciding which frequencies to utilize in US&R robot deploymentapplications, as will be discussed in Section 3.(a)(b)2.2 Excess Path Loss and RMS Delay SpreadWe also conducted measurements at several stationarypositions covering a very wide frequency band. These “excesspath-loss” measurements provide the received signal powerrelative to a direct-path signal over a frequency band. Whentransformed to the time domain, the wide frequency band yields ashort-time-duration pulse. This pulse can be used to study thenumber and duration of multipath reflections in an environment.Our synthetic-pulse, ultrawideband system is based on avector network analyzer (VNA). Our measurements coveredfrequencies from 25 MHz to 18 GHz. The post-processing andcalibration routines associated with it were developed at NIST[10]. In the synthetic-pulse system, the VNA acts as bothtransmitter and receiver. The transmitting section of the VNAsweeps over a wide range of frequencies a single frequency at atime. The transmitted signal is amplified and fed to a transmittingantenna. For this study, we used omnidirectional discone antennasfor frequencies between 25 MHz and 1.6 GHz, and directionalhorn-type transmitting and receiving antennas for frequenciesbetween 1 GHz and 18 GHz.(c)Figure 2: (a) Portal into the Hazel-Atlas mine tunnel. (b) 90 minside showing the bend depicted in Fig. 1(b) and the rough,sandy wall material. (c) Wood shoring approximately 150 m intothe tunnel. The robot we tested can be seen on the cart betweenthe railroad tracks.
#3#5#2#4#1#3#20#4#5-20Pavg [dBm]-4050 MHz-60-80-100-120-14050100150200250300350time [sec](a)#3#5#2#4#1#3#20#4#5-20Pavg [dBm]-40162 MHz-60-80-100-120-14050100150200250300time [sec](b)#3#5#2#4#1#3#20#4#5-20Pavg [dBm]-40448 MHzFigures 4 and 5 show measured excess path loss over a widefrequency band measured 50 m and 120 m, respectively, in theHazel-Atlas tunnel. Note that at 120 m, we have passed the rightangle turn in the tunnel. The top curve in each graph representsthe received power level, referenced to the calculated free-spacepath loss at that location. The bottom curve represents the noisefloor of the measurement system.Figure 4 shows that even in a line-of-sight conditionapproximately 50 m from the tunnel entrance, the spectrum of thereceived signal displays significant frequency dependence. Atfrequencies between 25 MHz and 1.6 GHz (Fig. 4(a)), the lossywaveguide effect is shown by the rapidly decreasing signal on theleft-hand side of the graph. We see that a carrier frequency higherthan approximately 700 MHz would suffer less loss compared tolower frequencies in this particular tunnel. Figure 4(b) showsfrequencies from 1 GHz to 18 GHz. In this case, we see frequencydependence in received power caused by strong reflections, asshown by the deep nulls and peaks in the top curve of Fig. 4(b).Once the receiving antenna turns the corner, see Fig. 5, thesignal takes on a more random variation with frequency sincetransmission consists of reflected signals only. For frequenciesfrom 25 MHz to 1.6 GHz (Fig. 5(a)), the received signal power isnear the noise floor of the receiver since the two curves almostoverlay. For the higher frequencies (Fig. 5(b)), we see that theaverage received signal level is relatively constant withfrequency, but the peaks and nulls are still significant.Finally, we present the RMS delay spread for the Hazel-Atlasmine tunnel in Table 1 for frequencies from 25 MHz to 1.6 GHzand 1 GHz to 18 GHz. We see that the shortest delay spreads arefound by use of the directional antennas. The complete set ofUWB excess-path-loss data is given in [4].-60-80-100-120-140We measured excess path loss every 20 m startingapproximately 10 m into the tunnel. The VNA was located at theHazel-Atlas portal. The transmitting antenna was located at theportal as well. The graphs show data starting from 0 Hz, howeverthe valid (calibrated) measurement range is stated for each graph.50100150200250time [sec](c)Figure 3: Received-power data in the Hazel-Atlas Mine forthree carrier frequencies: (a) 50 MHz, (b) 162 MHz, (c)448 MHz. In each case the #2 and #4 vertical dashed linescorrespond to the turn at 100 m: once on the way into thetunnel and once on the way out. The #3 dashed line representsthe end point at 200 m, shown in Fig. 1(b).The received signal was picked up over the air in the tunnelby the receiving antenna and was relayed back to the VNA via afiber-optic cable. The fiber-optic cable phase-locks the receivedsignal to the transmitted signal, enabling post processingreconstruction of the time-domain waveform associated with thereceived signal. The broad range of frequencies and time-domainrepresentation provide insight into the reflective multipath natureof the tunnel that cannot be captured by use of single-frequencymeasurements. The receive antenna must remain fixed duringeach measurement, so these tests are carried out at discretelocations, unlike the single-frequency tests.2.3 Tests of Robot CommunicationsWe carried out tests on a commercially available robot. Controland video were as-built for the commercial product. We used theomnidirectional antennas that came with the system for all tests inorder to assess the default capabilities of this robot.The robot we used is controlled with a 2.4 GHz spreadspectrum, frequency-hopping protocol, which was configured totransmit in the unlicensed 2.4 GHz industrial, scientific, andmedical (ISM) band. The control channel utilizes a modulationbandwidth of approximately 20 MHz. The output power of thebidirectional control link is nominally 500 mW.The robot transmits video by use of one of ten channelsbetween 1.7 GHz and 1.835 GHz. The robot we tested transmittedat 1.78 GHz by use of an analog modulation format that was nonbursted and non-frequency-agile. The video channel utilizedapproximately 6 MHz of modulation bandwidth. The outputpower was nominally 2 watts.
20Table 1: RMS Delay Spread for the Hazel-Atlas mine tunnel.Center column: Frequencies from 25 MHz to 1.6 GHz measuredwith omnidirectional antennas. Right column: Frequencies from1 GHz to 18 GHz measured with directional antennas. The grayshaded areas represent a non-line-of-sight propagation condition.Excess Path Loss (dB)0-20-40-60-80-100RMSd 50noise 10-120-14002004006008001000Frequency (MHz)120014001600Distance(m)(a)BD mineHazel AtlasRMSDelay Spread Delay SpreadHigh Freqs.Low 015.210.0Figure 4: Excess path loss measurements over a wide frequencyband carried out 50 m from the portal of the Hazel-Atlas mine.(a) 25 MHz to 1.5 GHz. (b) 1 GHz to 18 GHz.9015.28.410015.79.6110x7.5Excess Path Loss (dB)20-40-60-80-100D 50 mnoise-120-140030006000900012000Frequency (MHz)150001800020Excess Path Loss (dB)0-20-40-60-80-100-120d 120noise 10-14002004006008001000Frequency (MHz)120014001600(a)200(a)D 120 mnoise @100Excess Path Loss uency (MHz)1500018000(b)Figure 5: Excess path loss measurements carried out 120 m fromthe portal of the Hazel-Atlas mine. (a) 25 MHz to 1.6 GHz. (b)1 GHz to 18 GHz.The robot controller was located at the entrance to the tunnel,shown in Fig. 6. We positioned the robot inside the tunnel afterthe first bend in a non-line-of-site condition. The robot wasmoved through the tunnel on a cart, as shown in Fig. 2(c), so thatwe could check the control link even after video was lost. Every10 m, the video quality and control link were checked. Video wasrated qualitatively by the robot operator, and control was checkedby the ability of the operator to move the robot arm, and verifiedby a researcher in the tunnel. No attempt was made to providemore granularity in these tests. That is, we assumed that movingthe arm up was equivalent to moving it down or rotating it.Table 2 shows the results of our tests. We were able tocommunicate with the robot in a non-line-of-sight condition deepwithin the tunnel. This is consistent with the results of Fig. 5(b),which indicates that signals in the low gigahertz range shouldpropagate farther than those at lower frequencies.Table 2 also shows that control of the robot was possiblemuch deeper into the tunnel than we were able to receive video,even though the output power of the video channel is higher (2watts for video vs. 0.5 watt for control). However, a much higherdata rate is necessary to maintain high-quality video transmission,as opposed to the relatively small amount of data needed tocontrol the robot. Transmitting this large amount of data requiresa more stringent success rate than for the control channel;therefore, failure of the video before the control is notunexpected. The delay experienced in controlling the robot whenit was deep in the mine indicates packet loss and resend for errorcorrection under weak-signal conditions.
εR1 a 3 ε 1 b3 ε 1 RR , α TUNNEL 4.343λ2 1 1 4 4b , aα ROUGHNESS 4.343π 2 h 2 λ α TILT 4.343Figure 6: Robot operator positioned at the entrance to the HazelAtlas mine tunnel. The robot was operated in a non-line-of-sightcondition more than 100 m inside the tunnel.Table 2: Results of wireless communication link tests carried outinside Hazel-Atlas tunnel at Black Diamond Mines RegionalPark.Distance intunnel (m)Video qualityControl of arm(1.7 GHz)(2.4 GHz)100goodyes110goodyes120poor (intermittent)yes130poor (intermittent)yes(2b)π 2θ 2λ ,(2c)and λ is the wavelength, a is the width of the tunnel, b is theheight of the tunnel, and h is the roughness, all in meters. Otherparameters include εR, the dielectric constant of the rock walls,and θ, the angle of the tunnel-floor tilt in degrees.We set the parameters of the model to approximate the HazelAtlas tunnel, given below in Table 3. This model works well onlyfor frequencies will above the cut-off frequency, that is, forwavelengths significantly less than the dimensions of the tunnel[5, 6]. Hence, in Fig. 7 we show results for 448 MHz only. Atdistances around 80 m, the signal was able to propagate throughan air vent as well as through the tunnel, so the overall receivedsignal level increases. The good agreement between the measuredand modeled data led us to conclude that waveguiding plays asignificant role in radio propagation in these tunnels.Table 3: Parameters used in tunnel model.very pooryes150noneyes160nonedelay experienced170noneintermittent control180nonedelay experienced190nonedelay experienced0200nonedelay experienced-20205nonenonePath Gain (dB)1403. MODELED RESULTS3.1 Single-Frequency Path Gain ModelsTo study the extent of waveguiding in these tunnels, weimplemented an analytical model that simulates signalpropagation in tunnel environments having various physicalparameters [5, 6, 11]. Briefly, the model assumes a dominantEH11 mode in a lossy rectangular waveguide with the attenuationα in dB/m expressed for vertical polarization asα α TUNNEL α ROUGHNESS α TILT ,(1)where(2a)ParameterValueWidth2mHeight2mWall roughness0.3εr6tilt1 448 MHz model448 MHz measured-40-60-80-100-120204060Distance (m)80100Figure 7: Comparison of measured and modeled data for theHazel-Atlas tunnel. The carrier frequency is 448 MHz. Themodeled data simulate waveguide propagation for a waveguidewhose physical parameters approximate those of the tunnels.The model also lets us explore which frequencies may beoptimal for robot or other wireless communications in the tunnel.Figure 8 compares a number of commonly used emergency
approximately 120 m, the video quality degrades and the picturebecomes intermittent.0TunnelFlat EarthELM 0dBPath Gain (dB)responder frequencies as a function of distance within the tunnel.As discussed in [5, 6], the frequency-dependent behavior of thetunnel leads to a “sweet spot” in frequency. Below the sweet spot,signals do not propagate well, due to the effect of waveguidebelow-cutoff attenuation and wall loss. Above the sweet spot,free-space path loss (which increases with frequency) and αTILTdominate and signals do not propagate well. Again, models suchas these may enable a choice of appropriate frequency for US&Rrobot communications in tunnel environments. Note that theseresults are valid only for a tunnel with these dimensions, wallmaterials, and surface roughness. The curves would need to berecalculated for other types of Increasingdistanceinto tunnel-40Path Gain (dB)-60-80-100Frequencieswith lowestloss in tunnel-120-140-150050100Distance (m)150200Figure 9: Path gain curves for tunnel with a right-angle turnat 100 m (solid) and flat earth (dashed) environments. Thecurve labeled “ELM 0” indicates where the excess linkmargin calculation predicts loss of signal. As shown, thisoccurs approximately 150 m into the tunnel.-160-180-2003.2 Channel Capacity Model5001000150020002500Frequency (MHz)300035004000Figure 8: Path gain versus frequency for various distancesin a tunnel having physical characteristics similar to those ofthe Hazel-Atlas tunnel. Frequencies around approximately400 MHz to 1 GHz propagate better than either lower orhigher carrier frequencies.We also used the model to investigate the video performanceof the robot, described in Section 3.c. The frequency-hoppingcontrol channel would need to be modeled by use of othermethods, since it consists of several narrowband channelsfrequency hopping within a wide modulation bandwidth. In Fig.9, we plot the estimated path gain at a carrier frequency of1.78 GHz for the tunnel environment with a right-angle turn100 m from the receiver. We used the parameters in Table 3 forthe model. A path gain of 40 dBW was used as anapproximation for the turn in the tunnel at 100 m, based on workdone by Lee and Bertoni in [12]. We plot the flat earth path gainfor comparison.Figure 9 also shows the theoretically computed excess linkmargin (ELM). The ELM is the difference between the receivedsignal strength and the minimum receiver sensitivity. The receiversensitivity is determined by the thermal noise of the receiver andthe receiver’s front-end amplifier noise (5 dB, as a rule of thumb).The thermal noise is given by N kTB, where k is the Boltzmannconstant, T is the temperature in Kelvin, and B is the bandwidth ofthe receiver. In order for a wireless link to be maintained, theELM must be greater than zero dB.The ELM plotted in Fig. 9 agrees well with the measuredresults from Table 2, which show that the video completely dropsout between approximately 140 m and 150 m. Given thefluctuation in signal strength due to multipath fading in this tunnelenvironment, once the link margin drops below 10 dB atIn general, received RF power and bandwidth effectivelyplace an upper bound on the capacity of a communications link.The Shannon channel capacity theorem [13] can be used topredict the approximate maximum data rate for tunnelcommunications, even though the Shannon theorem is based onthe assumption of a Gaussian noise (low multipath) environment.For a given modulation bandwidth, the received signal powerrelative to the noise power determines the theoretical upper limiton the data rate (channel capacity). The Shannon capacitytheorem is given byC B log2 (1 S / N ) ,(3)where C is the channel capacity in bits/second, B is the channelbandwidth in hertz, S is the received signal power in watts, and Nis the measured noise power in watts. The capacity represented bythis equation is the upper limit, and in reality the capacity wouldbe difficult to attain with real hardware.For an analog transmission, Shannon’s limit gives us a way toestimate the channel capacity. The National Television SystemCommittee (NTSC) analog video channel that our robot used hasa video bandwidth of 4.2 MHz and a transmission rate ofapproximately 30 frames per second, where each frame consistsof 525 scanning lines, giving a line rate of 15.734 kHz [14]. Wecan place an upper bound on the amount of data that could betransmitted in each line by considering a typical implementationof NTSC, where each line is digitized into 768 pixels. This givesa digital scanning rate of approximately 12 MHz. Thespecification of 768 pixels per line is used in studio environments.We expect the potential channel capacity to be lower in theanalog transmission case.Figure 10 shows simulations of the Shannon limit for ourrobot’s 4.2 MHz video bandwidth, 1.78 GHz video channel. Table4 shows the distance into the tunnel where 12 Mb/sectransmission rate occurs assuming our maximum possible channel
capacity to be various fractions of the Shannon limit. Based onthis information, we would expect to encounter video problemssomewhere between 120 m and 130 m into the tunnel, whichTable 2 shows is indeed where we started to experience signaldegradation. Thus, we are able to form a rough estimate of thedistance into the tunnel where we expect the video to fail basedon a simple implementation of the Shannon theorem.200100% of Shannon Capacity90% of Shannon Capacity80% of Shannon Capacity70% of Shannon CapacityCapacity (Mb/sec)150as well as improved technology and system design for theemergency-responder community.5. ACKNOWLEDGEMENTThe authors thank Alex Bordetsky of the Naval PostgraduateSchool for facilitating the measurements during recentinteragency marine interdiction operation system tests. We alsothank Bill Dunlop, Steve MacLaren, and Dave Benzel ofLawrence Livermore National Laboratory for logistical andtechnical support. We are indebted to Frederick M. Remley fordetails on the NTSC video standard.6. REFERENCES[1]100[2]500050100Distance (m)150200[3]Figure 10: Channel capacity predicted by the Shannontheorem for a carrier frequency of 1.78 GHz and a videomodulation bandwidth of 4.2 MHz. At 120 m, where weexperienced intermittent video, 80 % of the Shannon limit is15.4 Mb/sec and 70 % of the Shannon limit is 13.5 Mb/sec.[4]Table 4: Distance into tunnel for a channel capacity ofapproximately 12 Mb/sec.[5]Fraction of Shannon capacity(%)Distance into tunnel(m)100133901298012270114[6][7][8][9]4. CONCLUSIONWe have presented measured data collected in a subterraneantunnel environment. Results showed waveguide-below-cutoff andwall attenuation effects. We saw frequency-dependent peaks andnulls in the channel due to strong multipath reflections andattenuation in the tunnel. In no
transmission and detection in tunnels at the Black Diamond Mines Regional Park near Antioch, California on March 19-21, 2007. Our goal was to investigate propagation channel characteristics that affect the reliability of wireless telemetry and control of Urban Search and Rescue (US&R) robots in tunnels and other weak-signal environments.
Bruksanvisning för bilstereo . Bruksanvisning for bilstereo . Instrukcja obsługi samochodowego odtwarzacza stereo . Operating Instructions for Car Stereo . 610-104 . SV . Bruksanvisning i original
10 tips och tricks för att lyckas med ert sap-projekt 20 SAPSANYTT 2/2015 De flesta projektledare känner säkert till Cobb’s paradox. Martin Cobb verkade som CIO för sekretariatet för Treasury Board of Canada 1995 då han ställde frågan
service i Norge och Finland drivs inom ramen för ett enskilt företag (NRK. 1 och Yleisradio), fin ns det i Sverige tre: Ett för tv (Sveriges Television , SVT ), ett för radio (Sveriges Radio , SR ) och ett för utbildnings program (Sveriges Utbildningsradio, UR, vilket till följd av sin begränsade storlek inte återfinns bland de 25 största
Hotell För hotell anges de tre klasserna A/B, C och D. Det betyder att den "normala" standarden C är acceptabel men att motiven för en högre standard är starka. Ljudklass C motsvarar de tidigare normkraven för hotell, ljudklass A/B motsvarar kraven för moderna hotell med hög standard och ljudklass D kan användas vid
LÄS NOGGRANT FÖLJANDE VILLKOR FÖR APPLE DEVELOPER PROGRAM LICENCE . Apple Developer Program License Agreement Syfte Du vill använda Apple-mjukvara (enligt definitionen nedan) för att utveckla en eller flera Applikationer (enligt definitionen nedan) för Apple-märkta produkter. . Applikationer som utvecklas för iOS-produkter, Apple .
Wireless# Guide to Wireless Communications Chapter 1 Introduction to Wireless Communications . Wireless Local Area Network (WLAN) - Extension of a wired LAN Connecting to it through a device called a wireless . network Each computer on the WLAN has a wireless network interface card (NIC) - With an antenna built into it .
Johnny's Selected Seeds. Moving a high tunnel can also help address some of the challenges of crop rotation in tunnels, where . Fayette County, KY Since 2011, Dr. Krista Jacobsen has been working with 30' x 72' mov-able tunnels that were designed and built at the UKHRF. There have been multiple design changes. All the movable tunnels .
New Jersey Student Learning Standards for English Language Arts . Page 1 of 12. Grade 4 . The standards define general, cross-disciplinary literacy expectations that must be met for students to be prepared to enter college and workforce training programs ready to succeed. The K–12 grade-specific standards define end-of-year expectations and a cumulative progression designed to enable .
