Wireless Sensor Networks lab 2: Collecting RSSI data to determine the best antenna and antenna location John Ritchison and Chris MoressiIntroduction There are many situations where a person would want to take measurements at remote locations. However, this creates the problem of retrieving the data. One solution is to have someone travel to the remote sites and manually retrieve the data. For most situations a better solution is to send the data electronically usually via wireless communication. This is generally known as telemetry and has been practiced for years. One of the IIHR laboratories is currently working on a problem of this nature. This lab has a grant to collect rain data for Johnson County and has rain gauges distributed across the County. At this time an employee regularly visits each site to collect the data; however, they are working on a system to automate this process using cellular modems to transmit the data. Before implementing this system they need to decide which antenna to use and where to place the antenna. In order to aid in this decision we traveled to six of these remote sites and made RSSI (Received Signal Strength Indicator) measurements. At each site we recorded multiple measurements for each antenna. In this report we go over how we made these measurements and discuss our results. Methods Before going out in the field and making RSSI measurements we developed a Visual Basic Script to work in conjunction with SecureCRT to automate the process of retrieving desired information from the cell modem. Specific documentation from SecureCRT was required to enable certain functionality in the script such as writing to the terminal, waiting for certain strings, and reading data from the terminal. Once this was figured out, we were able to use text parsing capabilities and graphic capabilities of Visual Basic to complete the script. The final version of the script: (1) sent an RSSI request to the cell modem via a serial connection, (2) read the complete response to the request, (3) parsed out relevant data, (4) and finally converted and displayed the data in dBm’s. Following the completion of the script and the testing of all the hardware we traveled to six sites to take RSSI measurements. For each site we collected a total of eight measurements. Six of the eight measurements were taken in the direct vicinity of the rain gauge and the other two measurements were taken at the road closest to the particular site. The measurements taken at the rain gauge were at three specified locations. They were the front corner (the front being the side with the locks), the middle (or center), and the back corner (the back being the side with the hinges). At these three locations we took measurement with two different antennas at two different elevations. The elevations used were relative to the rain gauge. One measurement was taken at a low level near the bottom of the rain gauge and another was taken at a high level approximately twice the height of the rain gauge. It was our intention to determine if one antenna outperformed the other, if there was a difference between low and high elevations and if there was a difference between locations on the platform.Results When we took each measurement, we recorded the value in an excel spreadsheet. Afterwards we created graphs for each location (front, middle and back) at each elevation. Each graph shows the values for both antennas side by side. This data and these graphs are shown here. It should be noted that the value -1 corresponds to “no signal.” Front Corner (Low)-120-100-80-60-40-2003 5 6 11 24 25Site IDdBmSm AntLrg Ant Front Corner Middle Back Corner Low High Low High Low High Road Site ID Sm Ant Lrg Ant Sm Ant Lrg Ant Sm AntLrg Ant Sm Ant Lrg Ant Sm AntLrg Ant Sm Ant Lrg Ant Sm AntLrg Ant 3 -99 -103 -101 -101 -97 -97 -103 -105 -99 -97 -101 -93 -93 -99 5 -1 -1 -105 -1 -1 -1 -1 -1 -1 -1 -1 -1 -101 -103 6 -89 -91 -81 -81 -89 -93 -77 -87 -89 -95 -83 -83 -79 -83 11 -83 -87 -73 -89 -85 -81 -75 -95 -75 -87 -79 -87 -75 -81 24 -95 -85 -73 -85 -83 -85 -83 -85 -97 -85 -89 -85 -77 -105 25 -93 -95 -87 -105 -97 -93 -85 -101 -91 -99 -85 -95 -85 -95Front Corner (High)-120-100-80-60-40-2003 5 6 11 24 25Site IDdBmSm AntLrg Ant Middle (Low)-120-100-80-60-40-2003 5 6 11 24 25Site IDdBmSm AntLrg Ant Middle (High)-120-100-80-60-40-2003 5 6 11 24 25Site IDdBmSm AntLrg AntBack Corner (Low)-120-100-80-60-40-2003 5 6 11 24 25Site IDdBmSm AntLrg Ant Back Corner (High)-120-100-80-60-40-2003 5 6 11 24 25Site IDdBmSm AntLrg Ant Road-120-100-80-60-40-2003 5 6 11 24 25Site IDdBmSm AntLrg AntDiscussion From the data which was collected it is clear that more often than not the small antenna had a better RSSI value than the large antenna; however, it is not clear, at first glance, whether or not this is statistically significant. This is why we performed a one-way ANOVA (ANalysis Of VAriance). Ideally, for this analysis, we wanted to use a large number of samples from the same location with the two antennas being the only variables. Unfortunately we did not take a large number of measurements for each location (front, middle and back) and elevation (low and high); so instead we used the measurements from all of the locations and elevations. Since the values at each site were similar and the variance due to different locations and elevations was small, this should be ok. From the analysis we received a P value of 0.001896. This means that we can reject the null hypothesis (there is no statistically significant difference between the large and small antennas) with a confidence of ~99.81%. Next we performed another one-way ANOVA comparing the low and high values. Since we had already determined that it is highly likely there is a significant difference between the two antennas, we only used the values for the small antenna in this analysis. Because of this and the fact that we could not use the road data (we did not take high and low measurements there) we had less than half as many data points. With only fifteen data points we received a P value of 0.009848. Here the null hypothesis is that there is no significant difference between placing the antenna low or high and we can reject this with a confidence of ~99.02%. The last analysis we performed had to do with differences in the RSSI values at different locations (front, middle and back). Due to the results from the first two analyses we only used values