Karel Malý

In fingerprinting-based localization systems, collecting time-series received signal strength (RSS) observations and averaging them is a common method for dealing with RSS fluctuation during the creation of the fingerprint database. However, in practice, these observations contain outliers caused by multipath propagation, interference, and hardware instability, which degrade the effectiveness of the averaging process. The modified Z-score method is widely adopted for outlier detection due to its simplicity and moderate accuracy. However, its performance depends strongly on the choice of scale estimator used to quantify data dispersion. A systematic evaluation of robust scale estimators under varying RSS observation sizes, with emphasis on detection reliability and computational cost, remains limited. This study evaluates six robust scale estimators within the modified Z-score framework, namely median absolute deviation (MAD), Sn, Qn, interquartile range (IQR), trimmed standard deviation (TSD), and Winsorized standard deviation (WSD). The goal is to identify a scale estimator that ensures reliable outlier detection while maintaining low to moderate computational cost for practical deployment in fingerprinting systems. Results show that, in terms of outlier detection performance, the IQR and Sn estimators achieve the lowest false detection rates across different observation sizes. In terms of computational efficiency, the MAD estimator provides the lowest processing time based on empirical runtime measurements. When both detection performance and computational efficiency are considered, the IQR estimator offers the most practical solution for preprocessing RSS time-series observations in fingerprint database creation for fingerprinting-based systems.

The projected clustering (PROCLUS) algorithm is a subspace clustering algorithm based on the k-medoids clustering approach. It is designed to address the challenges of irrelevant received signal strength (RSS) measurements in fingerprint vectors by focusing on meaningful subsets of RSS measurements from wireless APs, known as subspaces. Despite its robustness, its performance heavily depends on the chosen similarity metric, with Euclidean and Manhattan distances being the most common. While many researchers focus on modifying the algorithm to enhance performance, the impact of similarity metrics on clustering performance is often overlooked, despite its critical role in determining accuracy. As such, this paper evaluates the clustering performance of the PROCLUS algorithm using five distance-based similarity metrics—Euclidean, Manhattan, Cosine similarity, Canberra, and Chebyshev—across six experimentally generated fingerprint databases. It aims to identify the best similarity metric for maximizing clustering performance for each of the six fingerprint databases, with silhouette scores used as the clustering performance metric. Simulation results show that Cosine similarity is the most effective metric with the PROCLUS algorithm. It consistently produces clusters with the highest silhouette scores, all well above the 0.25 threshold and were 24% to 136% higher than the scores achieved using the other distance-based metrics across the tested fingerprint databases. The Canberra distance performed variably, while Euclidean and Manhattan distances were less reliable. The Chebyshev distance consistently underperformed in all the databases considered. The findings in this paper highlight the importance of choosing the appropriate similarity metric to perform clustering operations with the PROCLUS algorithm.

Fingerprint database clustering and localization using k‐medoids and k‐nearest neighbour (k‐NN) algorithms respectively typically use distance‐based fingerprint similarity metrics, with their performances dependent on the type of distance metric used. This paper proposes employing a pattern‐based metric, the context similarity coefficient (CSC), for both algorithms instead of traditional distance‐based metrics. The CSC accounts for fingerprint behaviour and the non‐linear relationships among fingerprints during the similarity measurement. The performance of both algorithms with the CSC as the similarity metric is evaluated on four publicly available fingerprint databases, using position root mean square error (RMSE) and silhouette score as performance metrics. These results are compared to those of the same algorithms using five distance‐based metrics: Euclidean, square Euclidean, Manhattan, cosine, and Chebyshev distances. The k‐medoids algorithm with CSC shows moderate clustering performance compared to the five distance‐based metrics considered. However, when combined with the k‐NN algorithm also using CSC, it achieves the highest localization accuracy, with at least a 29% improvement in position RMSE across all four databases. The results indicate that while k‐medoids with CSC may not create well‐separated clusters, combining it with the k‐NN algorithm with CSC as its similarity metric significantly enhances localization accuracy compared to distance‐based metrics.

The affinity propagation clustering (APC) algorithm is popular for fingerprint database clustering because it can cluster without pre-defining the number of clusters. However, the clustering performance of the APC algorithm heavily depends on the chosen fingerprint similarity metric, with distance-based metrics being the most commonly used. Despite its popularity, the APC algorithm lacks comprehensive research on how distance-based metrics affect clustering performance. This emphasizes the need for a better understanding of how these metrics influence its clustering performance, particularly in fingerprint databases. This paper investigates the impact of various distance-based fingerprint similarity metrics on the clustering performance of the APC algorithm. It identifies the best fingerprint similarity metric for optimal clustering performance for a given fingerprint database. The analysis is conducted across five experimentally generated online fingerprint databases, utilizing seven distance-based metrics: Euclidean, squared Euclidean, Manhattan, Spearman, cosine, Canberra, and Chebyshev distances. Using the silhouette score as the performance metric, the simulation results indicate that structural characteristics of the fingerprint database, such as the distribution of fingerprint vectors, play a key role in selecting the best fingerprint similarity metric. However, Euclidean and Manhattan distances are generally the preferable choices for use as fingerprint similarity metrics for the APC algorithm across most fingerprint databases, regardless of their structural characteristics. It is recommended that other factors, such as computational intensity and the presence or absence of outliers, be considered alongside the structural characteristics of the fingerprint database when choosing the appropriate fingerprint similarity metric for maximum clustering performance.

In fingerprint-based localization systems, clustering fingerprint databases is a proposed technique for improving localization accuracy while reducing localization time. Among various clustering algorithms, density-based spatial clustering of applications with noise (DBSCAN) stands out for its robustness to outliers and ability to accommodate fingerprint databases of various shapes. However, the clustering performance of the DBSCAN algorithm is heavily influenced by the type of similarity measure metric used, with most researchers using distance-based metrics. This paper aims to enhance DBSCAN clustering by using a pattern-based metric known as the context similarity coefficient (CSC) instead of distance-based metrics. The CSC metric examines received signal strength (RSS) measurement patterns that form fingerprint vectors and assesses both linear and non-linear relationships between these vectors to determine similarity. Four publicly available fingerprint databases were used to evaluate the clustering performance with silhouette scores as a performance metric. The performance of the DBSCAN algorithm with the CSC metric is determined and compared to Euclidean and Manhattan distances as similarity measure metrics. Simulation results indicate that achieving good clustering performance with the DBSCAN algorithm requires generating three or fewer clusters. The proposed CSC metric demonstrated the best clustering performance in two of four fingerprint databases and the second-best in another. However, computational complexity comparisons reveal that the CSC metric is highly computationally intensive and is suggested to be used on small to medium-sized fingerprint databases generated using an odd number of wireless APs deployed in a non-uniform or non-grid-like distribution.

Distance-based metrics are the most common fingerprint similarity metrics used in fingerprint database clustering and localization processes in a fingerprint-based localization system. In this paper, however, a less common but promising pattern-based fingerprint similarity metric is proposed as an alternative to the distance-base metric. The proposed fingerprint similarity metric is based on an inverse weight (IW) normalization of the context similarity coefficient (CSC)-based similarity metric measure. The clustering and localization performance of the fingerprint-based localization system with the proposed IW-CSC-based fingerprint similarity metric is determined and compared to the square Euclidean, Manhattan, and cosine distance-based metrics. The k-means algorithm with a k-means++ cluster initialization process is considered for fingerprint database clustering, while the k-nearest neighbor (k-NN) algorithm is considered for localization. Based on the four fingerprint databases considered, the proposed IW-CSC-based metric has the slowest localization time with moderate clustering performance. However, it has the best localization performance, which is at least 52% higher than the localization performances of the three distance-base metrics considered. The proposed IW-CSC-based metric is recommended as an alternative to the distance-base metric only when improved localization performance is the primary objective of the fingerprint-based localization system. It is also recommended for use in small to medium-sized fingerprint databases for clustering and localization.

The modified Z-score (mZ-score) method has been used to detect outliers in time series received signal strength (RSS) observations. Its performance is dependent on the scale estimator used, and each has advantages and disadvantages over the others. One approach to developing a scale estimator that combines the advantages of two or more scale estimators is through scale estimator hybridization. In this paper, the outlier detection performance of a mZ-score method with different hybridization approaches for Sn and median absolute deviation (MAD) scale estimators is determined and analysed. Three different hybrid scale estimators are identified, namely weighted, maximum, and average hybrid scale estimators. The performance of the mZ-score method using the three different hybrid scale estimators is determined using three experimentally generated and publicly available time-series RSS datasets. Based on the simulation results, the weighted hybrid scale estimator results in the best outlier detection performance amongst the three hybrid scale estimators. When compared to the mean-shift-based outlier detection (MOD) technique, the k-means clustering-based technique, and the density-based spatial clustering (DBSCAN) technique, the mZ-score method with the weighted hybrid scale estimator performs better with little or no false alarm and false negative detections.

The performance of fingerprint-based indoor wireless localization systems (IWL-Ss) can be enhanced using fingerprint clustering. The localization performance of clustered fingerprint-based IWL-Ss is affected by several factors, including choosing the most optimal initial parameters and the appropriate fingerprint similarity measurement metric. The problem of choosing the best initial parameter is solved by using the affinity propagation clustering (APC) algorithm in this paper, which automatically calculates the number of clusters and cluster centroid vectors. However, the choice of fingerprint similarity measure and the selection of the best cluster centroid when there are multiple potential cluster centroids limit the performance of the APC algorithm. To address this issue, this paper proposes modifying the conventional APC (c-APC) algorithm, which will be referred to as the "m-APC algorithm." The context similarity coefficient (CSC) fingerprint similarity measure replaces the distance-based fingerprint similarity measure used by the c-APC algorithm. Furthermore, the cluster centroids that are generated automatically are replaced by the centroid that is obtained by averaging all fingerprints within a cluster. Using the k-NN localization algorithm and four online fingerprint databases, the performance of the m-APC+CSC algorithm is determined and compared to the c-APC algorithm using cosine, Euclidean, and Shepard distances as fingerprint similarity measures. Based on simulation results, the m-APC algorithm reduced the position root mean square error (RMSE) and mean absolute error (MAE) by about 12% and 8%, respectively, when compared to the c-APC algorithm when both used the CSC as a fingerprint similarity measure. Furthermore, the m-APC+CSC algorithm achieved an 8% and 9%, respectively, position RMSE and MAE reduction over the c-APC algorithm using cosine, Euclidean, and Shepard distances as similarity measurements. The m-APC+CSC algorithm should, however, be used on a reasonably sized fingerprint database with at least four wireless access points (APs) for better localization performance.

Decision structures have been proposed in a number of contexts an knowledge-based software engineering as an important mechanism for recording and reasoning about the information needed to make decisions during the software lifecycle. We apply decision structures to a new domain, namely that of the business rules of an enterprise. Given the dynamic nature of business rules, decision structures are seen to be an appropriate framework to record and evolve business rules. The paper describes how to combine decision structures and business rules within a conceptual modeling framework. An architecture is presented that addresses business rules throughout the operational lifetime of the systems they govern.