1 Uninort..

Notas del editor

• #4: Las pérdidas de espacio libre son aquellas causadas por la propagación en el aire, asumen línea de vista total entre las antenas, y no tienen en cuenta pérdidas por obstrucciones. Ya que la potencia es radiada sobre la superficie de una esfera cuya área es proporcional a su radio, la reducción en el nivel de señal es proporcional al cuadrado de la distancia del transmisor. Estas expresiones son válidas para campos lejanos. En zonas cercanas a la antena (campos cercanos) los campos electromagnéticos son muy complejos y estas expresiones no son válidas. The receive power is calculated by taking the transmitted power (in dBm) and subtracting the path loss, adding the antenna gains and subtracting any other equipment losses that may exist. In mobile systems the gain of the mobile antenna is generally considered to be 0dB.
• #6: When an electromagnetic wave passes from a medium with one refractive index to a medium with a different refractive index, the ray is ‘bent’ as shown in the diagram
• #7: When an electromagnetic wave encounters an obstruction, or partial obstruction, the obstruction causes a shadow. The shadow has blurred edges as shown in the diagram. The extent of the blurring increases with wavelength. This phenomenon is known as diffraction. The strength of a radio signal may, then, be influenced by obstacles which do not obstruct the line of sight as well as those that do.
• #8: The free space model does not adequately describe propagation in the mobile environment. Propagation takes place above, and close to, the earth’s surface. This can be modeled simply as shown in the diagram. For the sake of simplicity the earth is considered to be perfectly flat and smooth (plane) and a perfect conductor. This results in a reflection with zero loss. There is a 180 degree phase change at the reflection point with a further small phase change due to the difference in path length between the direct ray and the reflected ray. The exact phase difference is dependent upon the path geometry i.e.. the path length and the antenna heights. The loss increases with an increase in path length and decreases with an increase in antenna heights. The loss is also independent of frequency. Calculations show that the path loss is proportional to 40log(d) when using the plane earth model, rather than 20log(d) in the free space model, and to -20log(ht) and -20log(hr), where ht and hr are the antenna heights.
• #9: The earth’s atmosphere changes with altitude. The most significant change as far as electrical properties are concerned is the fall in water vapor content with increase in altitude. This results in the refractive index of the atmosphere falling with increase in altitude. An electromagnetic wave that is launched horizontally into the atmosphere, because of the curvature of the earth, will pass through progressively higher atmospheric layers that have progressively lower refractive indices. This results in the wave bending back towards the earth as shown in the diagram. This bending of the wave results in it traveling beyond the horizon. A convenient way of describing this is to assume that the wave has traveled in a straight line and that the radius of the earth has changed so moving the horizon. We therefore define an “effective earth radius”. This may be calculated by multiplying the real earth radius by an “effective earth radius coefficient” which, in a normal atmosphere, may have a value from 2/3 to 4/3.
• #10: As mentioned previously, the atmospheric refraction of radio waves causes them to “bend” around the curvature of the earth thereby extending the radio horizon. The effect can be described by modifying the real earth radius by a factor known as the “effective earth radius coefficient”. The effective earth radius coefficient, for a normal atmosphere, will fall between the values of 2/3 and 4/3. This allows the path taken by the radio wave to be represented by a straight line in order to simplify calculations. The bulge of the earth changes the position of the reflection point with respect to the antennas. This results in a concept, shown in the diagram, of effective antenna height. This effective antenna height is used in loss calculations. It should be noted that the introduction of hills and other terrain features will also change the effective antenna height and hence the path loss.
• #11: Diffraction calculations can be very complex depending upon the shape of the obstruction over which the wave is propagating. The simplest type of obstruction to deal with is a knife edge as shown in the diagram. The loss caused by the obstruction can be calculated from the antenna heights, the obstruction height and the distance of the obstruction from each antenna. The diffraction loss will change depending upon the shape of the obstruction. It is possible to add a correction factor that is dependent upon the radius of the top of the obstruction to model this. It is also common practice at VHF and UHF frequencies to consider any terrain feature to be a knife edge. As frequencies increase and the wavelength gets shorter, so a given obstruction begins to look more rounded and less like a knife edge.
• #12: This curve shows the variation of diffraction loss with clearance over an obstruction. The quantity (v) is an indication of the degree of clearance or obstruction. A +ve value indicates a clearance and a -ve value indicates an obstruction.
• #13: Propagation may take place over several obstructions. When this happens there are a number of different ways of dealing with the problem. It is possible to treat every knife edge as an additional diffraction loss or to construct an equivalent knife edge. Summing diffraction losses tends to lead to a pessimistic result and there are methods that add weights to each loss in an attempt to correct this. Constructing and equivalent obstruction and treating the problem as a single knife edge leads to a rather optimistic result but is fairly quick and simple to do.
• #14: So far we have examined the theory of propagation as applied to the mobile environment. The following sections of this course reveal the real complexity of the problem of predicting the behavior of radio propagation in a mobile environment. The problem is so complex that it is never possible to gather sufficient data to produce an absolute answer but only to predict that we may expect a result within a certain range of values.
• #15: A real atmosphere is unstable with its characteristics changing with altitude and time. The effect on short paths at low frequencies as currently used in mobile systems is very small and not usually considered. Ducting can take place in the troposphere at VHF and UHF frequencies and this can result in severe long range interference problems. This effect is usually temporary (but not in some parts of the world) and there is little that can be done about it.
• #16: The reflections that have been considered so far are termed specular (mirror like). The behavior of these reflections from a smooth, perfectly conducting surface is much the same as light being reflected from a perfect mirror. In reality the reflective surfaces that exist are rough and much more like a mirror that has been rubbed with wire wool. The result is a reflection that does not concentrate the reflected power in one direction but scatters it over a wide area. There may even be some back scatter. The reflected signal in any one direction is therefore of much lower amplitude. The actual amplitude will depend upon the roughness of the surface. Power reflected from a wide area of a rough surface will reach the receive antenna. This results in many signals , with many different amplitudes and phases, adding together and making it very difficult to calculate what exactly will happen at any one point. For the purposes of characterizing the channel so as to be able to determine signal distortion and to test the performance of real equipment, a 16 ray model is required.
• #17: The diagram on the left shows a reflection from the side of a building. The magnitude of these reflections depend upon the construction of the building and the angle of incidence of the signal. Modern metalised glass windows make very good reflectors. This effect can be observed by studying visible reflections at different angles. The diagram on the right shows that the problem is a three dimensional one. Radio waves propagate down city streets as though they were canyons with highly reflective sides. Although this effect may seem to be troublesome at first it can be and is used to provide micro-cell coverage in high traffic city areas where antennas can be placed below roof top level.
• #18: The penetration of buildings by radio waves depends upon the construction of the building, the frequency in use and the angle of arrival of the wave. In general, for macro-cell designs, penetration losses are lower on the higher floors for concrete and steel frames buildings. Losses decrease as the frequency increases because the spaces between the steel members of the building become larger compared to the wavelength. This is not true of buildings using large areas of metalised glass where losses increase with frequency. Propagation within the building also depends upon construction but losses are generally proportional to 40log(d). It is not unreasonable to expect increased losses in the order of 20dB for building penetration at the frequencies used in mobile systems.
• #19: Real obstructions rarely conform to the nice theoretical idea of a knife edge. To deal with real obstruction, they may be simplified into a series of multiple knife edges as shown. These multiple knife edges may then be treated as described previously. It is also possible to draw lines (shown in red on the slide) in such a way that an equivalent single knife edge can be constructed. All these methods have their drawbacks and it is only experience and particular circumstances that will lead to really accurate predictions.
• #20: Time dispersion is another problem relating to multiple paths to the Rx antenna of either an MS or BTS. However, in contrast to Rayleigh fading, the reflected signal comes from an object far away from the Rx antenna. Time dispersion causes consecutive bits to interfere with each other making it difficult for the receiver to determine which symbol is the correct one. An example of this is shown in the figure below where the sequence 1, 0 is sent from the BTS. If the reflected signal arrives one bit time after the direct signal, then the receiver detects a 1 from the reflected wave at the same time it detects a 0 from the direct wave. The symbol 1 interferes with the symbol 0 and the MS does not know which one is correct. ADAPTIVE EQUALIZATION Adaptive equalization is a solution specifically designed to counteract the problem of time dispersion. It works as follows: 1. A set of predefined known bit patterns exist, known as training sequences. These are known to the BTS and the MS (programmed at manufacture). The BTS instructs the MS to include one of these in its transmissions to the BTS. 2. The MS includes the training sequence (shown in the figure as “S”) in its transmissions to the BTS. However, due to the problems over the radio path, some bits may be distorted. 3. The BTS receives the transmission from the MS and examines the training sequence within it. The BTS compares the received training sequence with the training sequence which it had instructed the MS to use. If there are differences between the two, it can be assumed that the problems in the radio path affected these bits must have had a similar affect on the non-training sequence bits. 4. The BTS begins a process in which it uses its knowledge of what happened the training sequence to correct the other bits of the transmission.
• #21: Each MS on a call is allocated a time slot on a TDMA frame. This is an amount of time during which the MS transmits information to the BTS. The information must also arrive at the BTS within that time slot. The time alignment problem occurs when part of the information transmitted by an MS does not arrive within the allocated time slot. Instead, that part may arrive during the next time slot, and may interfere with information from another MS using that other time slot. Time alignment is caused by a large distance between the MS and the BTS. Effectively, the signal cannot travel over the large distance within the given time. TIMING ADVANCE Timing advance is a solution specifically designed to counteract the problem of time alignment. It works by instructing the misaligned MS to transmit its burst earlier than it normally would.
• #23: Propagation over a plane earth would give us a fall in received level proportional to 40log(d) (shown in red on the slide). The effects of terrain shadowing give us a signal level with a log normal distribution about the theoretical level and a standard deviation of about 8dB. This is shown in blue on the slide. The effect of reflections is to give the signal level a Raleigh distribution. The depth of these fades can be over 40dB. The three effects sum to give a signal level that behaves as shown in black on the slide. Fast fading on LIne of Sight is Known as Riccian fading, which is lower than Raleigh.
• #24: The complexity of the propagation effects described and their dependence upon the accurate location and description of all features makes it impossible to say that the received level at a particular point of a signal coming from a particular transmitter is an absolute level. All we are able to do is to make statements such as “there is a 98% probability that the received level will be within 8dB of an absolute value. To be able to make this statement we must have a mathematical model that is accurate enough to do this. Although the models that are used for mobile predictions have a sound theoretical base they are all empirically derived. That is to say that somebody has made a large number of field measurements of receive levels and, from a theoretical base, has constructed a mathematical model, the output of which matches those measurements. Many people have done this and published their findings. The result is a number of models that each work well for a given set of general circumstances.
• #25: Las estadísticas nos pueden ayudar a ampliar de manera efectiva lo que sabemos de la física y lo que observamos en las mediciones para predecir niveles de señal en lugares en donde no se pueden hacer mediciones.
• #26: Las estadísticas nos pueden ayudar a ampliar de manera efectiva lo que sabemos de la física y lo que observamos en las mediciones para predecir niveles de señal en lugares en donde no se pueden hacer mediciones.
• #27: Según los visto hasta ahora, se deben combinar todas las herramientas de tal forma que se puedan obtener resultados acertados (niveles probabilísticos y de certidumbre) en el desarrollo de los modelos.
• #28: Los modelos analíticos simples son útiles para calcular respuestas y niveles de señal específicos en ciertos puntos, pero no permiten hacer una generalización para hacer el diseño del sistema. Ejemplo no puedo diseñar solo con los cálculos de espacio libre (y el efecto de los edificios?, y de los lagos? Y de las vías? Son preguntas que estos modelos no responden)
• #29: Son modelos empíricos. Estos modelos son apropiados para realizar el cálculo de estaciones base que cubrirán un área y para el diseño temprano de una red. Sin embargo, no dan una predicción exacta en el cálculo de puntos aislados pues siguen siendo sólo estadísticas.
• #30: La cobertura de cada sitio es calculada haciendo análisis de propagación y niveles de señal para miles de ubicaciones, distribuidas de manera uniforme en el área que deseamos cubrir.
• #31: Este tipo de modelos caracterizan las fluctuaciones de señal a nivel “microscópico” en ubicaciones determinadas dando niveles de probabilidad de confiabilidad de servicio
• #32: Los modelos derivados empíricamente corresponden a estudios desarrollados por grupos de personas o instituciones interesadas en profundizar y perfeccionar las técnicas de diseño de sistemas inalámbricos.
• #33: Models are generally based on the principle that the level (measured in dB) falls in a linear fashion with distance from the transmitter. This is represented by a term in the model of Ka.log(d) where Ka is the slope. At some distance from the transmitter the level is set to a fixed value. This takes the form of a “magic number” and is known as the intercept. An offset may be applied for effective base station antenna height or mobile effective antenna height all along the path. “ Local” offsets may be applied to the model at different points to reflect the effects of different clutter types at different points along the path or the effects of a diffracted path i.e.. shadowing by terrain or other obstructions.
• #34: Recopilaciones hechas en Tokio Japón durante mas de 7 meses por Yoshihisa Okumura en 1968. The Okumura - Hata model has been derived from prediction curves obtained from experimental data. The model is suitable for use in the design of land mobile systems operating in the range of 100 to 1800 MHz. This is, in fact, a modification of the original model produced Hata. The base station effective antenna height should be in the range from 30m to 200m. Mobile antenna heights should be in the range of 1m to 10m and the distance between base station and mobile should be greater than 1Km. This model is obviously a generalized model and requires a number of correction factors to suit it to specific applications. Los términos son: 69.55 una constante 26.16log (f) las pérdidas se incrementan con la frecuencia 13.82 log (hb) las pérdidas se disminuyen 14dB con un incremento de la altura de la RBS por un factor de 10. (44.9-6.55log (hb))log(d) las pérdidas se incrementan en un múltiplo del logaritmo de la distanca. Ese múltiplo es una pendiente de atenuación que es determinada por la altura hb. Si es pequeña como 10 m la pérdida es 38 dB/década, si es 100m las pérdidas son 32 dB/década, a 1000m las pérdidas son 25 dB / década. El modelo de okumura es el padre de los modelos Okumura – Hata y Cost 231
• #35: A correction factor is calculated and applied that describes the variation in terrain heights over the path. A fine correction factor may be applied also. This gives more accurate information about the variation in peak heights over the path. If line of sight exists between the base station and the mobile and inclination correction factor is applied which describes the mean slope between the two. A correction factor may be applied if there is water beside the mobile or the base station. Correction factors are applied depending upon the type of clutter found at the mobile location. A correction factor is applied so as to describe diffraction losses along the path. This may be done for single or multiple knife edges. A correction factor is applied depending on mobile antenna height to compensate for the effect of ground proximity on antenna performance. El modelo de Okumura Hata es ampliamente utilizado en áreas urbanas muy pobladas pero debido a lo factores de corrección de clutter es utilizado ampliamente.
• #36: This model is used where canyon propagation exists in city streets and is for use between 800Mhz and 2000MHz. Information about road orientation with respect to the line of transmission is required as is information about street width and building heights The model is limited to base station heights of 50m, mobile heights of 1m to 3m and path lengths of 20m to 5Km. Presenta unos niveles de precisión muy buenos con unos errores medios de predicción de aprox 3 dB y desviacióon estándar de 4-8 dB cuando la altura de la antena de la radio base está por encima de la altura del nivel medio del techo, pero empeora cuando la antena está por debajo del nivel medio del techo.
• #37: COST (Cooperación Europea en el campo de la Investigación Científica y técnica) ha hecho un gran esfuerzo desde el inicio de los desarrollos de sistemas móviles en Europa para elaborar métodos poderosos de predicción. Adicionalmente propone un la definición de un marco de trabajo común para comparar el rendimiento de los modelos propuestos. Al igual que el Okumura Hata sus términos pueden ser explicados
• #38: Microcells are generally of the order of 500m to 1000m radius with antennas situated below rooftop level, typically at street lighting level. A mobile turning a street corner can suffer sudden losses of 20dB to 30dB within a few feet. The macro-cell models can be out by as much as 20dB to 30dB if used for micro-cell predictions. Propagation falls roughly into two categories, line of sight and non line of sight. Non-line-of-sight paths may best be split in to two section having two different loss/distance slopes.
• #39: It is possible that a change in terrain type or some other effect may cause a change in the loss slope at some distance from the base station. This can be modeled by choosing to use different slope and intercept values after some particular break point.
• #40: Before any predictions can be made these values must be assigned and the values must reflect the effects of the real world. The only way to do this is to carry out a number of test drives to make real world measurements and then select values for all the variables so that the predictions reflect, as accurately as possible, the measured values. This need not be done for every cell, but for representative cells in each area.
• #41: The diagram represents a number of signal level measurements taken at various points within the coverage area of a cell. In practice there would be over a thousand of these measurements. It is possible to draw a straight line through this plot that will show the underlying slope of the level/distance characteristic. To test the accuracy of the line that has been drawn it is necessary to calculate the error at every measurement point and hence a mean error. If the line that had been drawn was the blue one instead of the red one there is obviously an error. If the mean error is calculated, because there are both positive and negative errors, it will come to zero. To test the slope, therefore, the RMS error must be calculated.
• #42: The slope of the line is now fixed. It is possible to move the line up or down on the plot. If this is done and the mean error, between the line and the actual measurements, is calculated it is possible to place the line so that there is close to zero mean error. The diagram shows a red line with the correct offset and a blue line with an incorrect offset. It is now possible to mark the plot at a fixed distance from the base station and to obtain a value in dBm for the intercept point. This point is shown marked in green on the diagram. The slope and intercept values have now been calculated and may be used in the propagation model.
• #43: The local variations in level may be due to clutter at the mobile location. In this slide the samples have been color coded to indicate the type of clutter present at each sample site. This helps in deciding what sort of value to assign to each sort of clutter. Having assigned clutter values, the model must be run and its predictions compared with the real measurements. The calculation of mean errors in different types of clutter and the standard deviation of errors enables these values to be fine tuned. There is also an overall clutter weighting to be assigned.
• #44: Diffracted paths are detected by observing the terrain heights along the path from the base station to the mobile. The Software reports on any diffracted paths it finds and calculates the predicted losses. The losses are calculated using single or multiple knife edges, depending upon circumstances , and an overall weighting factor is applied. This weighting factor will take account of the general shape of hills, soil conditions and other factors in the local area. the weighting factor is adjusted until the error between the model and the real measurements is minimized.
• #45: The diagram shows what is meant by effective antenna height. There are several ways in which effective antenna height can be determined, there are four options. The antenna height above ground may be added to the difference in ground height at the base station and the mobile (providing that the mobile is below the base station antenna height). The average terrain height for the entire coverage area may be calculated and the height of the antenna above this value used. The slope of the terrain may be calculated and an effective antenna height obtained as shown in the diagram. The average height along the path from the base station to the mobile may be calculated and the antenna height above this value used.
• #46: When antennas are close to the ground, in terms of wavelengths, there is an effect which alters the characteristics of the antenna and hence its gain. A factor is therefore included in the general model to account for this.
• #47: This course has, so far, taken a very simplistic view of propagation. Where reflections occur, propagation has been shown as a “ray” and reflections as occurring at a single point. This is not true. Reflection occurs over a large area the size of which depends upon wavelength and the distance from the transmitter and receiver. For this reason it is more accurate to consider the effect of clutter for some distance in front of the mobile (between it and the base station.) The effect of the clutter will not remain constant over this area and so some variable scaling function is required. Clutter loss calculation then consist of a clutter type factor, a clutter scaling coefficient and a clutter weighting function. Clutter height information may be used along the path from the base station to the mobile to better calculate obstruction losses. Clutter separation factors are used to separate the mobile from the surrounding clutter to prevent it from being “swamped”.