Multipath Detection and Mitigation by Means of a MEMS Based Pressure Sensor for Low-Cost Systems

摘要

MEMS barometric pressure sensors have evolved tremendously over the past years and have emerged into the consumer domain, such as cell phone platforms and personal navigation devices. The main objective on the use of the barometric pressure sensor in conjunction with GPS is aimed at enabling the pedestrian navigation use case, e.g. floor detection and further improvement of the altitude accuracy. As the pedestrian use case predominantly operates in a challenging multipath (MP) environment, e.g. urban canyon, the performance of cell phone based GPS receivers is often limited by MP. This paper does not aim at proposing to use the pressure sensor measurement to improve the vertical guidance accuracy, but to enhance the GPS position engine by leveraging the barometric pressure sensor measurement to detect and mitigate MP. First, the algorithm framework is established to enhance MP detection by means of simulated data. It is proposed to engage the pressure sensor as a measurement, which operates in the ranging domain of the GPS system (a quasi Pseudolite). Subsequently, by formulating subsets of the fullset satellite solution, and quantifying the dispersion of the normalized Sum Squared Error (SSE), it is proposed to perform MP detection. In order to do so, the derivation of the test-statistic to compare the subset solutions with the fullset solution is a necessary step to formulate a decision criterion. As the pressure sensor measurement is characterized by a different error model compared to a GPS ranging measurement, the test statistic needs to incorporate a normalization technique to transform the inhomogeneous sample space. Key factor for acceptable performance is an accurate calibration of the pressure sensor. After successfully detecting a MP distorted measurement, the specific satellite ranging measurement is removed from the solution. The validation of the proposed algorithm is performed by using a simulation environment as well as real data. Specifically a consumer grade GPS chipset is employed, as well as a low-cost MEMS based pressure sensor. The results will quantify the merit of detecting and mitigating MP on a low-cost platform using a pressure sensor aided GPS receiver as opposed to an unaided receiver. The quantification will be done in terms of probability of miss detection and false alert. The focus is aimed at the capability of detecting and rejecting MP, as opposed to the increased accuracy in terms of vertical positioning (as this has already been addressed in various papers).