Unsupervised Learning Techniques for Trilateration: From Theory to Android APP Implementation

Abstract

Because the characteristics of wireless propagation channels (especially indoor channels) are too diverse and complex, the distance estimation strategy of range-based positioning techniques should adaptively change depending on the environment. In this paper, we study unsupervised learning techniques that efficiently do this without human intervention. As users simply move around an area of interest with mobile devices, the proposed method autonomously learns the characteristics of the surrounding environments and changes the ranging strategy accordingly. To this end, we use either model-based or neural network (NN)-based ranging modules for estimating the distance from neighboring anchor nodes, calculate the position of the devices using trilateration techniques, and define cost functions that indirectly evaluate the accuracy of the ranging module based on the trilateration results. Moreover, by assigning a unique trainable variable to each device, the proposed method is also able to compensate for different characteristics between devices without ground truth data. The performance of the proposed method is verified with a real-time location tracking application using received signal strength (RSS) measurements from conventional Wi-Fi access points (APs) or round trip time (RTT) measurements from APs that support the fine timing measurement (FTM) protocol. In cases where a model-based ranging module is used, the proposed method closely achieves the benchmark performance, which perfectly optimizes all the trainable variables on the test data. If NNs are adopted in the ranging module, the proposed method even outperforms the benchmark and achieves an average positioning accuracy of up to 2.397 m using RSS measurements, and up to 1.547 m using RTT measurements under the 40 MHz bandwidth configuration. INDEX TERMS Fine timing measurement (FTM), neural network, positioning, trilateration, unsupervised learning.

Unsupervised Learning Technique to Obtain the Coordinates of Wi-Fi Access Points

“…According to the experiment results in [13], [14] and this work, the uncalibrated FTM protocol provides distorted distance measurements of the true distances. This is due to the timing offset of the FTM packets, cable length, hardware offset, or even site-specific factors 1 .…”

“…For the same reason, the extended Kalman filter (EKF) is also compatible with the proposed method, because it consists of matrix operations. We follow the EKF procedures introduced in [14].…”

“…Irrespective of which trilateration algorithm is used, we can obtain the estimated coordinates of the device and define cost functions that indirectly evaluate the accuracy of the parameters [14]. The most important cost function is the geometric cost function that is defined by…”

“…This protocol measures the distance between two nodes based on the round trip time (RTT). Although the performance of the FTM protocol still depends on many factors such as the presence of a line-of-sight (LOS) path, obstacles in the LOS path, the transmission bandwidth, hardware calibration, and so on, it potentially provides more accurate ranging results compared to RSS [12]- [14].…”

“…Moreover, given that users can go anywhere in the area and connect multiple APs in the middle, the proposed method can be applied even if some unknown APs are isolated from others. To this end, the cost functions introduced in [14] are used to evaluate the geometric validity of the relationship between the coordinates of anchor nodes and distance measurements. In addition to estimate the coordinates of unknown APs, we also focus on the calibration of the FTM protocol.…”