Weather Stations-Assisted Barometric Altimeter
for Android: Interpolation Techniques
for Improved Accuracy

[pdf-embedder url=”https://wellapets.com/wp-content/uploads/2019/06/Weather-Stations-Assisted-Barometric-Altimeter.pdf” title=”Weather Stations-Assisted Barometric Altimeter”]

Abstract—The increasing availability of smartphones equipped with barometric sensors enables their use as barometric altimeters.
However, an accurate measurement requires updated information about the temperature and the mean sea level pressure at the location of the measurement. With this aim, this work proposes, analyzes, and compares spatial interpolation methods applied to atmospheric pressure and temperature reports from nearby reference weather stations in order to derive reference values to be used by altimeters. As a proof-of-concept, we implemented an Android Altimeter application. Its validation was performed by a set field experiments. Preliminary results show the potential of the proposed methods for the accuracy of smartphone-based barometric altimeters.
Index Terms—altimeter, barometric sensor, interpolation methods, smoothing techniques, weather stations 

INTRODUCTION
Location based services for three-dimensional space positioning usually resort to GPS to determine the elevation but, even if the horizontal accuracy is acceptable for the majority of usage scenarios, the vertical accuracy is not sufficiently reliable for applications requiring accurate altitude information.
Altimeters are traditionally used in activities such as aircraft flights, rescue operations, military operations, skydiving, and climbing. Despite the existence of other types of altimeters such as the sonic and radar altimeters, the pressure-based altimeter still is largely used by aircrafts.
The increasing availability of smartphones equipped with barometric sensors enables their use as barometric altimeters.
However, the solutions that we propose in this work are generic, being applicable both to smartphones and to special purpose hardware (if equipped with barometric sensor, computation, and communication capabilities).
An accurate measurement requires updated information about the temperature and the mean sea level pressure at the location of the measurement. For this purpose, mechanical barometric altimeters [1] require user intervention either by adjusting the “barometer setting” with the mean sea level pressure reported by a weather station, or by setting the altimeter to a known altitude [2]. On other hand, smartphone applications [3] usually retrieve the “barometer setting” from a web server in order to automatically adjust the altimeter. However, the poor accuracy of altitude measurements of these applications led us to investigate the source of these inaccuracies, and how to overcome them.
A prevalent problem is the barometer accuracy. Since these sensors are less accurate than the ones certified for professional use, the values measured pressure values tend to be noisy. This problem can be easily mitigated by using a simple low-pass filter [4].
Since, the density of weather stations is sparse, the accuracy of the altitude calculation depends on the distance to the stations and on the current sea surface level atmospheric pressure gradient.
To our knowledge, existing applications neglect the fact that atmospheric data is collected at scattered locations, thus resorting to reports from the nearest station (polygonal method [5]), which is often distant from the point of interest for the altitude calculation. This introduces an error in the altitude estimation that could be avoided interpolating the pressure values of the nearby weather stations with appropriate methods [5], [6].
Weather stations typically broadcast weather reports on a hourly or half-hourly basis. The atmospheric pressure values are often rounded down before transmission, having a 1 hPa resolution. The error introduced is usually lower than 9 meters (at standard temperature). However, in the time span between weather reports, the atmospheric conditions may change considerably.
One solution to this problem, still not used in this work, is to request a couple of recent past weather reports in order to perform temporal extrapolation (e.g., polynomial or Kriging 1D [5]), to estimate the trend of the variation. An alternative solution could be the use of forecast models of
atmospheric pressure variation.
Aiming at improving the accuracy of barometric based altitude calculations, we propose, analyze, and compare spatial interpolation methods applied to atmospheric pressure and temperature reports from nearby reference weather stations.
The interpolated values are then used to adjust the altimeter setting. As a proof-of-concept, we implemented an Altimeter application for Android smartphones equipped with barometric sensors.
We present the following main contributions:
1) We propose and analyze spatial interpolation methods to estimate the reference atmospheric pressure and temperature to be used by the barometric altimeter;
2) As a proof-of-concept, we implemented an Android Altimeter application. However, the proposed solution is suitable to be used in special purpose Altimeter devices;

3) We evaluate and compare the proposed interpolation methods for the altitude calculation using real experiments.
The remainder of this paper is organized as follows: Section II reviews existing smartphone applications for altitude measurement. Section III discusses how the atmospheric pressure can be used to derive the altitude of some measurement point above mean sea level. Section IV discusses candidate spatial interpolation techniques for the derivation of the atmospheric pressure and temperature references. Section V presents the main features of the proof-of concept Altimeter Android application implemented in this work. Section VI evaluates and compares the proposed solutions via experimental field tests. Finally, Section VII presents the main conclusions of this work, and discusses future research directions.