Learning to Improve WLAN Indoor Positioning Accuracy Based on DBSCAN-KRF Algorithm From RSS Fingerprint Data

Abstract

WLAN-based indoor positioning algorithm has the characteristics of simple layout and low price, and it has gradually become a hotspot in both academia and industry. However, due to the poor stability of Wi-Fi signals, received signal strength (RSS) fingerprints of some adjacent reference positions are difficult to evaluate similarity when utilizing traditional distance-based calculation methods. By clustering these RSS fingerprints into one region, the commonly utilized KNN algorithm in the past cannot achieve accurate positioning in the region. For this, we introduce a concept of the insensitive region of the RSS fingerprint and a new algorithm named DBSCAN-KRF. This algorithm can delete noise sample and detect insensitive region, then, different methods are selected to achieve indoor positioning by judging the region of the estimated fingerprint sample, the KNN algorithm is selected when the region is sensitive, and random forest algorithm is selected when the region is insensitive. The experimental results show that the DBSCAN-KRF algorithm is superior while compared with other alternative indoor positioning algorithms. INDEX TERMS WLAN indoor position, control and optimization, machine learning, DBSCAN-KRF algorithm, fingerprint data.

High Precision Weighted Optimum K-Nearest Neighbors Algorithm for Indoor Visible Light Positioning Applications

“…To statistically evaluate the performance of our proposed algorithm and other existing methods, we apply Friedman test, which is used to rank the algorithms' performance. The test statistic suggested by Friedman (14) and F distribution by Iman and Davenport (15) are as follows [33,34] According to the null hypothesis, all the compared algorithms have the same rank. Therefore, the null hypothesis is rejected when the ranks of all algorithms are not equivalent.…”

“…In the first step, we computed the average ranks of each considering algorithm (i.e., RKNN, ROKNN, RWKNN, and RWOKNN) and the mean rank Rmean of algorithms. Then, we computed the values of TF (14), FF (15), and the critical value of F distribution for significant level of 0.05. With 4 algorithms and 51 data points, the critical value of F distribution is 2.79, which is greater than the mean rank Rmean.…”

“…Based on these data, the distance from the transmitter to the receiver can be estimated by converting the RSS data into the length between LEDs, or estimated locations can be inferred by combining RSS data and the fingerprinting method [2,14]. The RSS is a popular, simple approach that provides acceptable positioning error for indoor positioning applications [13]; This approach is also applied for other types of signals such as WIFI [15,16], Bluetooth [17,18], Lorawan [13], and Zigbee [19,20]. In the VLPS field, recent research based on the RSS platform have continued to improve [9,11,21].…”

Improving DBSCAN for Indoor Positioning Using Wi-Fi Radio Maps in Wearable and IoT Devices

“…DBSCAN [11]- [13] is a clustering method used to split the radio map into high-density and low-density clusters, dividing it into n non-overlapping reduced radio maps. Then, in the operational phase, the search has two steps.…”

Change Detection of RSSI Fingerprint Pattern for Indoor Positioning System

“…To evaluate the change detection performance, four different scenarios to provoke signal pattern variation are considered. The positioning is executed by the k-NN (Nearest Neighbor) method [36]. From the experimental results, it is found that the similarity metric decreases corresponding to the increment of the positioning error.…”