WAVE Communication-based V2I Channel Modeling

2.1 WAVE Communication Standard

WAVE communication was standardized in the IEEE 802.11p, which was written to modify the PHY and MAC layers of IEEE 802.11a. The latter is the wireless LAN (WLAN) standard, which corresponds to wireless connection between vehicles and roadside stations.

Table 1 shows the major characteristics of IEEE 802.11a and IEEE 802.11p. Unlike IEEE 802.11a, which uses the 5-GHz Industry Science Medical (ISM) band, IEEE 802.11p uses the independent frequency band of 5.850–5.925 GHz to reduce co-channel interference caused by band collision. Furthermore, it guarantees enhanced multipath mitigation by reducing the channel bandwidth of 20 MHz to 10 MHz and supports a transmission rate of 3–27 Mbps depending on the modulation method.

Table 1:

Comparison of standard IEEE 802.11a and 802.11p

2.2 Modeling of wireless channel

Channel modeling of the V2I system for public roads is crucial in implementing highly reliable WAVE communication for effective intelligent transport system (ITS) services. Propagation wave channel modeling refers to a graphical representation of the attenuating pattern of the propagation wave radiating from the transmitting terminal to the receiving terminal according to the environment. The propagation model can be interpreted by using both large-scale fading and small-scale fading. Large-scale fading refers to propagation analysis considering path loss due to the increasing distance between the transmitting terminal and the receiving terminal as well as the shadowing that occurs because of the obstacles between them. Channel modeling is complete once small-scale fading is analyzed by considering multi- path mitigation based on large-scale fading. In this paper, a large-scale fading model, which is the basis of the WAVE communication channel model, is constructed.

The log-distance model [5], which is the basis of the path-loss model, is expressed as:

where P_t denotes the measured transmission power of the transmitting terminal and P_r denotes the measured reception power of the receiving terminal. The signal attenuation can be determined with respect to distance d. Equation (2) represents Friis’ free space propagation model [6], which can derive the propagation attenuation size in free space with respect to antenna gain. G_tr denotes the antenna gain of each transmitter and receiver, λ denotes the wavelength of the WAVE communication band, and L_system denotes the system loss.

Friis’ free space propagation model, which assumes an ideal environment, is difficult to apply to an actual road environment where the ground causes reflection of radio waves. Thus, a two-ray ground-reflection model [6] that considers the largest reflected waves of the ground is primarily used in this study. This model is used to derive Equation (3) by calculating the largest reflected waves from the ground in the attenuation model while considering the heights of the transmitter and receiver, where h_t is the height of the transmitter and h_r is the height of the receiver.

The Friis model can be applied if the distance is smaller than the crossover distance d_x, at which the attenuation characteristics between the transmitter and the receiver in free space change. Meanwhile, because the two-ray ground-reflection model can better represent inter-vehicle communication on the road, it can be applied favorably if the distance is longer than d_x. This crossover distance is defined by the following Equation (4):

The test in the present study used WAVE communication in the 5.9-GHz band with transmitter and receiver heights of 3 m and 1.5 m respectively. Thus, d_x ≈ 1109 m. In other words, if a vehicle is 1109 m away from the station, an attenuation of four times the distance occurs. Furthermore, the reference distance d₀ of the far field can be included in the log-distance model of Equation (5). Here, the far-field reference distance of the WAVE communication is calculated to be 10 m [7] or 1 m [8], which are the WLAN standards. In this study, a distance of 10 m is applied.

In Equation (5), k is the attenuation coefficient, for which 2.0 is used for free space and 4.0 or higher is used for an urban environment. When the dual-slope model [7], which can select different models for different reference distances, is applied for this case, Equation (6) is obtained. For this dual-slope model, the log-distance model of free space with a coefficient k is applied if the distance is smaller than the reference distance, and the two-ray ground-reflection log-distance model is applied with the new coefficient k₂ otherwise. Therefore, a dual slope model that can express long-range propagation characteristics can be derived by using the two-ray ground-reflection model based on reference distance, as shown in Equation (6).

If the log-normal model is applied using X_σ, which is the zero-mean normally distributed random variable with standard deviation from Equation (6), we obtain the following Equation (7):

If the log-normal model is divided by the reference distances d₀ and d_x, and applied to Equation (6), it results in Equation (8). Here, for X₁ and X₂, the probability distribution based on the dispersion at each distance is applied.

The WAVE channel model is analyzed based on the log-normal of Equation (8).

3.1 Measurement Environment

To design a model optimized for South Korea domestic environment, we measured the WAVE signals on both the Yeoju experience road and on the Seoul test road in which the V2I infrastructure had been constructed [9][10]. Figure 1 shows the location and surrounding environment of the roadside stations on the experience road, located south of the Yeoju Interchange. The triangles in Figure 1 represent the stations themselves, with red indicating an unused station and green indicating an active one. As shown in Figure 1, five stations were installed on this road at about 1-km intervals. A section of about 1.5 km from the transmitter was measured from station no. 2. The Yeoju experience road was constructed alongside the southward expressway coming out of the Yeoju Interchange and contains almost no obstacles because it is not in use as a public road.

Figure 1:

A map of the Yeoju experience road and the locations of its roadside stations

Figure 2 shows the station locations of the test road in Gyeonggi Province and their surrounding environment. The triangles in Figure 2 indicate the locations of roadside stations, red and green once again indicating active and inactive stations respectively. Seven stations were installed between Seoul Interchange and Singal Interchange with about 1-km intervals between them, and only the third and the seventh stations were activated in order to minimize interference from the radio waves of adjacent stations. To create an environment that was more similar to an actual road-driving situation, the WAVE signals were measured while driving an actual vehicle.

Figure 2:

A map of the test road and the locations of its roadside stations

The test road was a fully operational expressway with a typical road environment that was being used by numerous vehicles. Therefore, it was possible to measure various WAVE signals that had been distorted by obstacles in other lanes, such as large vehicles.

Because the experience road and the test road had different road environments, we were able to obtain various data points, such as the propagation attenuation characteristics in free space and some signals that had been distorted by obstacles in the two environments. A propagation channel model for a public road was created by using the results of an area with almost no obstacles, and the reliability of this model was verified by comparing it to the results of the actual road situation when obstacles were included.

3.2 Measurement Method

We used a 5.9-GHz antenna and a spectrum analyzer to collect the WAVE signals transmitted from the V2I base-stations installed along both the Yeoju experience road and the Seoul test road section. The existing WAVE communication stations were used as transmitters. To design a basic large-scale fading model for the Yeoju experience road, the WAVE signals were measured at least 10 times, in order to remove the small-scale fading phenomena at each location, and at intervals of 10 m, in order to obtain sufficiently dense data points, over a total distance of 1.5 km, using the stations on a road section containing the smallest number of fixed obstacles. Figure 3 shows the second roadside station installed on the experience road. The directional receiving antenna and the three-directional transmitting antenna in the red circle were set at heights of 1.5 m and 3 m, respectively.

Figure 3:

A roadside station used for WAVE communication

Because the test road section was an expressway in full operation, and it was impossible to make measurements under static conditions, we collected the measured WAVE signals while driving. Due to the noise created by obstacles such as the vehicles in other lanes, which had not been present on the experience road, this model was verified by using both the model that had been derived from the Yeoju experience road as well as the signals measured on the test road. Figure 4 shows the vehicle that was used for measurement. The antenna in the red circle is the omnidirectional antenna. Table 2 shows the major antenna factor characteristics of the transmitter and the receivers used in both Yeoju and Seoul. The vehicle carrying these measurement devices travelled back and forth between the Seoul Interchange and the Singal Interchange five times at the highway speed limit of 80 km/h.

Figure 4:

The vehicle used for measuring WAVE communication

Table 2:

Antenna factor characteristics of the transmitter and the receivers

4.1 Measurement Results

In this section, the data measured on the Yeoju experience road and the Seoul test road is analyzed. The WAVE communication frequency band was assigned to 75 MHz, but only one channel band of frequency 10 MHz with a central frequency of 5.860 GHz was used to minimize the interference of the surrounding frequencies in this measurement. Furthermore, the output value of the signal that was received by the transmitting antenna from the station communication model was 20 dBm, the gain was 12 dBi, and the system loss was 5 dB. Measurements were made in 10-m increments during movement, and the first measurement point was about 10 m away from the station. In summary, index 1 was 10 m and index 100 was 1 km from the station [11][12].

Figure 5 shows the maximum number of signals measured at index 3 for each frequency. The x-axis represents the frequency of the WAVE signals and the y-axis represents the WAVE reception signal strength. Figure 6 shows the maximum signal strength for each frequency at index 100. The x-axis represents the frequency of the measured WAVE signals and the y-axis represents the WAVE reception signal strength. From Figure 5 and Figure 6, we can see that WAVE signals at around 5.860 GHz were received from the 10 MHz band, and that a –60.8 dBm signal is inputted in a 30-m section. Furthermore, a signal of around –101.1 dBm was received in a 1-km section. Similar to Figure 5 and Figure 6, Figure 7 shows the mean value of the center 10 MHz band at 5.860 GHz by distance.

Figure 5:

The signal strength of the WAVE band of index 3

Figure 6:

The signal strength of the WAVE band of index 100

Figure 7:

The average signal strength according to position

Figure 7 shows the measured results of the 5.860 GHz WAVE signals according to the index (in 10-m intervals) from the roadside station. The x-axis represents the position index, and measurements were made in 10-m intervals during movement from the first measurement point. The y-axis represents the strengths of the received WAVE signals measured at the receiving terminal. It can be seen that the measured values decrease in logarithmic form with distance, so the log-distance model mentioned in Section 2 can be applied. The measured strengths of the received signals between indexes 1–3 was weak because, as shown in Figure 3, the sign in front of the transmitter antenna was obscuring the path between the transmitter and receiver antennas. Furthermore, slightly low WAVE signals were recorded in the sections about 200 m and 600 m from the first measurement due to the fixed obstacles obscuring the transmitter. The strength of the received WAVE signals decreased as the distance from the first measurement grew, and it fell to –100 dBm or lower after around the 1 km mark.

4.2 WAVE Communication Channel Modeling

Channel modeling was performed based on the measured results of the WAVE signals. The dual-slope log-distance model, which can take propagation characteristics according to the reference distance into partial account by using a two-ray ground-reflection model that can consider the reflection coefficient of the ground, was applied. The advantages of this method were its ability to analyze communications for transmission and reception distances larger than those that other communication methods allow for and its ability to reflect the signal characteristics that vary by distance in free space. Figure 8 shows the log-distance model created when coefficients of 2.0-2.8 were applied to determine the optimum coefficient k₁ from Equation (5). Here, the x-axis represents the position index and the y-axis represents the WAVE reception signal strength. The red line in this graph denotes the crossover distance, which is the point at which the propagation attenuation characteristics change. The curve created by k₁ is applied only to the graph sections to the left of the red line. Because the measured data represents the maximum value at each index position, it only has to correspond to the model curve on top of the measured data graph according to the coefficient. Therefore, the most appropriate attenuation graph for the measured data is obtained at k₁ = 2.4.

Figure 8:

The log-distance model based on k1 and the origin data

Because a two-ray ground-reflection model was applied, Friis’ model is not applicable after the crossover distance, as described in Chapter 2. Therefore, a new coefficient k₂ was required. The new model was applied to the graph sections after the distance d_x based on d_x ≈ 1109 m, which was derived from Equation (4). As shown in Figure 9, the coefficient was varied between 2.8 and 3.2 in order to determine the value of k₂ . The x-axis in this graph represents the position index and the y-axis represents the WAVE reception signal strength. The most appropriate parameter value for the coefficient k₂ was determined to be 3.0.

Figure 9:

The two-ray ground-reflection model based on k2 and the origin data

Finally, the parameter values required for the modeling process were determined, as shown in Equation (9). The final modeling equation can be produced using these values.

Equation (10) shows the model derived from the above process in this study.