Project Description

As a leader in electromagnetic propagation modeling, Dekalb Telecom receives many projects that require accurate propagation modeling. One such case that Dekalb Telecom’s participated in was from the Georgia DA office to solve a high profile political murder.

The project was setup as a competition divided into two phases to narrow the number of firms involved in the final project. Phase 1 was a demonstration of our propagation modeling technology to narrow down the participating firms based on the accuracy of our propagation modeling. The Georgia DA provided us with geographical information and a few of received power measurements so that our model could estimate the path-loss effects of terrain, roads, buildings, etc. Once our propagation model determined the path-loss exponents, our model was tested by the Georgia DA by testing our model on four locations. The estimates from this test would then be compared to actual measured values. The firms producing the most accurate propagation models would advance onto phase 2 of the competition.

Phase 2 would involve work on the actual murder case. Since the Georgia DA subpoenaed the cell phone records of the primary subjects, it was our job to predict the probability that the subjects were near the vicinity of the victim during the time of the murder and that both subjects were involved in the murder. The firm that produces the most accurate phase 2 analysis would be awarded the contract for the project. 

Phase 1 Details

For phase 1, the Georgia DA provided us with data about the received power in different locations, clutter data, elevation data and the base station specifications, such as its azimuth, EIRP, transmitter location, frequency, scaling factor and height. This data was then inserted into our propagation model which utilizes free space attenuation, clutter diffraction and terrain diffraction theories to estimate path-loss exponents that can later be used to estimate signal strength at any point on the map. For more information on our propagation modeling technology, please look at our Technology page. For the results from our propagation modeling, please look at our Technology Demonstration page.

Phase 2 Details

In November 2000, a new sheriff was elected and was murdered one month later. Police believe that the incumbent sheriff was involved in the murder and identify Patrick C and Paul S as the primary suspects. Please view the Police Report for more information about the case. Both subjects work for Burger King, shown as a flag on Figure 1, and they both claim to be working all night. Since the police department subpoenaed the cell phone records of each suspect, we know which cell towers handled each of the suspects’ calls at specific times. The cell towers involved with handling the suspects’ calls the day of the murder are shown in Figure 1.

Figure 1. Cell towers that handled the suspects' calls the night of the murder.

Since cell phone antennas serve 120° sectors and have a maximum range of a 3-mile radius, as plotted on Figure 1, the general vicinity of the suspects can be located at specific times. These are "best" case specifications and most cell phone towers have a less ideal range. Figure 2 shows the antenna coverage pattern for the Allgon 7184.15 cell phone tower used for towers 214, 222, 223, and 224. Figure 3 shows the antenna coverage pattern for the Allgon 7184.13 cell phone tower used for tower 041. Since the maximum range of the antenna is 3 miles, Figures 2 and 3 show that this range is achieved only in the center of the antenna's sector, and the range drops off off-center. This helps to show that our approximation in Figure 1 is a "best" case approximation and can place the subjects in near the cell phone towers with high certainty.

Figure 2. Allgon 7184.15 Coverage Range.	Figure 3. Allgon 7184.13 Coverage Range.

The location of the subjects was further reduced, as shown in Figure 1, by an estimation of which roads the subjects likely were travelling on and by the areas of interest to the case. Using such techniques to narrow the possible location of the subjects' at specific times, Figure 1 shows the most likely location of the subjects shaded in different colors.

As can be seen in Figure 1, the suspects’ cell phone records clearly contradict their claim that they were at Burger King all night. In addition, our analysis shows that both suspects were within a 1 mile radius of the victims home between approximately 8 PM and 9 PM. Police testing using an actual cell phone also confirms that tower 214 does handle calls from the victim's home. The police believe that the suspects were snooping around the victim's home during this time. Our analysis also shows that the subjects were most likely at Burger King between 9 PM and 10 PM. Of most interest to the case is that Mr. S was in the blue shaded area in Figure 1 at 10:44 PM and Mr. C was there 10:30 PM and 12:01 AM. The victim's home is separated from the blue shaded area by approximately 5 miles and 10 minutes, with the assumption that the suspects were at the furthest point away from the victim's home within the reach of tower 222. According to our analysis, it is plausible that Mr. C could have left the blue shaded area, arrived at the victims home, committed the murder and returned back to the blue shaded area between 10:30 PM and 12:01 AM. It is also plausible that Mr. S left the shaded blue area at 10:44 PM and made it to the victim's house before 11:19 PM, the approximate time of the murder.