A Low Complexity System Based on Multiple Weighted Decision Trees for Indoor Localization

Abstract

Indoor position estimation has become an attractive research topic due to growing interest in location-aware services. Nevertheless, satisfying solutions have not been found with the considerations of both accuracy and system complexity. From the perspective of lightweight mobile devices, they are extremely important characteristics, because both the processor power and energy availability are limited. Hence, an indoor localization system with high computational complexity can cause complete battery drain within a few hours. In our research, we use a data mining technique named boosting to develop a localization system based on multiple weighted decision trees to predict the device location, since it has high accuracy and low computational complexity. The localization system is built using a dataset from sensor fusion, which combines the strength of radio signals from different wireless local area network access points and device orientation information from a digital compass built-in mobile device, so that extra sensors are unnecessary. Experimental results indicate that the proposed system leads to substantial improvements on computational complexity over the widely-used traditional fingerprinting methods, and it has a better accuracy than they have.

Indoor Location Sensing with Invariant Wi-Fi Received Signal Strength Fingerprinting

“…On the other hand, the proposed method excludes non-invariant Wi-Fi signals such that α becomes around 0.3, as shown in Figure 11. Note that, recent work by Sánchez-Rodríguez et al [27] shows that a computational complexity of O (1) is feasible in terms of floating point computation based on multiple weighted decision trees although the decision trees may grow with O(α m × n ).…”

“…Recently, research in WiFi-based indoor localization explored its feasibility using smartphones with various techniques [11,17,18,19,20,21,22,23,24,25,26,27,28]. In choosing the Wi-Fi signals to be used to uniquely classify a location, different authors used different approaches in their methods.…”

Indoor Localization Method Based on Regional Division with IFCM

“…In [31], a convolutional neural network was proposed that uses only one access point channel state information, and achieves an average positioning error of 1.36 m, but has a high training complexity. On the contrary, a low computational complexity model was proposed in [32], which shows a positioning error of 2.1 m. In this model, RSS and direction information are combined by the C4.5-based AdaBoost algorithm to improve the accuracy of indoor positioning. Crowdsourcing [33,34] can provide sufficient fingerprint updates for indoor LBS.…”

SDN-based Handover in Future WLAN

“…Although, several position determination schemes [24][25][26][27] has been introduced to estimate position in the wireless network. In this work to estimate the position of MS, we adopt the position estimation method that is divided into two phases [26] are described as follows.…”