More than two-third of 911 calls come from wireless devices, which are not associated with a specific location in the way that landline phone lines are. Since 2014, the Federal Communications Commission has been in the process of laying out rules to increase the accuracy of wireless 911 callers’ locations, so that they can be more quickly found by first responders.
As the FCC sought to set its accuracy requirements, it needed information about the state of capabilities and what level of accuracy was possible, particularly in vertical location to locate callers in multi-story buildings, also known as Z-axis information. This is a new piece of information for wireless operators to deliver and for 911 call centers to accept and utilize. The FCC required the nationwide wireless providers at the time to establish an independent indoor test bed to evaluate the performance of available technologies to provide Z-axis information, and put out calls to third parties to test their technology: the 911 Location Technologies Test Bed.
What was Stage Z testing? The Stage Z test campaign was the first round of testing, conducted in 2018, using CTIA’s 911 Location Test Bed (administered by Further Enterprise Solutions [FES] and managed by the Alliance for Telecommunications Industry Standards [ATIS]). Its goal was to provide a “rigorous, transparent process to evaluate the accuracy and overall assessment of Z-axis technology based on standard testing methodologies.”
The purpose of the testing was not to compare competing vendors’ solutions, but to demonstrate the state of the technology and to help the FCC develop a metric on how accurate it could reasonably require indoor location to be. Ultimately, the agency decided on a metric of ±3 meters.
Who participated in Stage Z testing? NextNav and Polaris Wireless participated, both of them testing technologies that rely on barometric pressure sensor information from mobile wireless handsets to determine an estimated altitude of an indoor wireless 911 call.
What did Stage Z testing consist of? According to the testing report, testing was conducted in Atlanta, Georgia; San Francisco, California; and Chicago, Illinois, with testing taking place in a variety of buildings including high-rise residential and commercial buildings in dense urban, urban, suburban and even some rural areas. In each building tested, several test points were chosen to represent different barometric pressure environments within the building, and the different types of areas from which a 911 call might be received. The report notes that Chicago was chosen as the third test location specifically to test the effects of more extreme weather conditions and potential impacts of fluctuating indoor/outdoor temperature and pressure differences on the barometric pressure sensor technologies.
Testing occurred in a total of 48 buildings, with 312 test points among those buildings. A total of 30 smartphone devices of 12 distinct model types were used to produce more than 100,000 location fixes. Apple devices were able to be tested by only one of the vendors at the time. The testing also consisted only of determining altitude, with no effort to determine an actual building floor or what supplemental information might be needed to convert that altitude figure to a floor level.
What were some of the conclusions of the testing report? The “single most important message of the testing,” according to the report, is that active background calibration of the barometric pressure sensors was “essential to achieve consistent and reasonable Z-axis estimation measurements for indoor wireless 911 calls due to mobile wireless handset biases that significantly affect the accuracy of barometric pressure-based estimation systems.” The report also said that the testing didn’t sufficiently answer questions about real-world deployment or scalability. The testing was incomplete, with NextNav’s tech not tested in Chicago or rural areas at the time, and Polaris’ tech not able to be tested on Apple devices. Nonetheless, both vendors were able to achieve accuracy within five vertical meters for 80% of their fixes. NextNav achieved vertical error of 1.8 meters or less for 80% of its fixes, and Polaris fixes had a vertical error of 4.8 meters or less. It should be noted that this is 2018 data; both companies have further developed their technology. Polaris asserted in a February 2020 filing with the FCC that not only can it meet the ±3 meter standard, but that it has “achieved sub-2 [meter] accuracy in several non-public test environments.”
Have any other companies used the Z-axis test bed? The test bed held a second round of testing, Stage Za; and according to its website, is now in a third round of testing (Stage Zb). Google has used the testbed, but successfully petitioned the FCC to have its test results kept confidential. According to an FCC fact sheet, the testbed’s Stage Za testing in 2019 evaluated Google’s Android-based Emergency Location Service. As reported by CTIA to the FCC, the Google ELS achieved ± 3 meter accuracy for more than half of calls in the test bed, and exceeded the 80th percentile metric in one morphology — which would not meet the FCC’s requirements for providing information for 80% of calls. (The various morphologies were dense urban, urban, suburban, or rural settings.) Apple was expected to test its Hybridized Emergency Location solution in the Test Bed’s Stage Zb testing campaign, which was originally slated to begin field testing in October 2020 but was delayed due to the Covid-19 pandemic. According to the testbed’s published schedule, Stage Zb testing will evaluate indoor location accuracy of a device calling 911 in 55 buildings across the four morphologies; the start date was pushed back to December 1, 2021.
Looking for more information about indoor and vertical location information for wireless 911 callers? Keep an eye out for RCR Wireless News’ upcoming report on the topic.
