Due to substantial research in indoor positioning a vast amount of location technologies and algorithms are available to enable various applications. However, it is challenging knowing which positioning system is optimal, or sufficient, for a specific application---not only when considering development and maintenance costs, but also the potential impact of the positioning system's performance on the respective application's performance.

The cooperative, invisible non-clinical work of hospital orderlies is often overlooked. It consists foremost of transferring patients between hospital departments. As the overall efficiency of the hospital is highly dependent on the coordination of the work of orderlies, this study investigates the coordination changes in orderlies' work practices in connection to the implementation of a workflow application at the hospital. By applying a mixed methods approach (both qualitative and quantitative studies), this paper calls for attention to the changes in orderlies' coordination activities while moving from a manual and centralized form to a semi-automatic and decentralized approach after the introduction of the workflow application. We highlight a set of cross-boundary (spatial and organizational) information-sharing breakdowns and the challenges of orderlies in maintaining local and global coordination. We also present design recommendations for future design of coordination tools to support orderlies' work practices.

Indoor Positioning based on WiFi has received extensive focus from research communities, owing its promises of ubiquitous positioning indoors for a variety of applications. The majority of the existing research on WiFi positioning centers around the improvement of positioning accuracy. However, to this end there is missing a unanimous evaluation methodology appropriately reflecting criteria relevant in real-world use. Assessing and advancing the state-of-the-art is often hindered by incomparable evaluation settings, and in many cases a missing rigor in describing evaluation methodology, metrics, and data collection, which is often decoupled from realistic use scenarios and their diversity. This leads to major challenges—and often disappointments—when implementing positioning methods in the hope of the promised accuracies. We believe that this holds especially in large real world environments, as only few extensive experimental evaluations have been reported in such environments. This paper presents a comparative evaluation of several indoor positioning techniques set at a large university hospital spanning 160, 000m2—an environment where several location- and context-aware logistic applications are in daily use. The positioning methods utilize measurements of signal strengths using existing WiFi infrastructure, which eases deployment and maintenance. We identify meaningful key metrics which describe different aspects of the methods' performance. Using these metrics, we furthermore report on experiences with implementing and utilizing indoor positioning solutions in a highly diverse environment, in which building types and materials, as well as building use differ across the complex. Correspondingly, the evaluation data we use is gathered at different building complex parts, days, and daytimes, and both at static locations as well as traveling within the building complex. Our results illustrate and quantify the challenges and breakdowns in transferring performance results from a small controlled setting, such as a small office environment, to a large dynamic building complex.

Being aware of activities of co-workers is a basic and vital mechanism for efficient work in highly distributed work settings. Thus, automatic recognition of the task phases the mobile workers are currently (or have been) in has many applications, e.g., efficient coordination of tasks by visualizing co-workers' task progress, automatic notifications based on context awareness, and record filing of task statuses and completions. This paper presents methods to sense and detect highly mobile workers' tasks phases in large building complexes. Large building complexes restrict the technologies available for sensing and recognizing the activities and task phases the workers currently perform as such technologies have to be easily deployable and maintainable at a large scale. The methods presented in this paper consist of features that utilize data from sensing systems which are common in large-scale indoor work environments, namely from a WiFi infrastructure providing coarse grained indoor positioning, from inertial sensors in the workers' mobile phones, and from a task management system yielding information about the scheduled tasks' start and end locations. The methods presented have low requirements on the accuracy of the indoor positioning, and thus come with low deployment and maintenance effort in real-world settings. We evaluated the proposed methods in a large hospital complex, where the highly mobile workers were recruited among the non-clinical workforce. The evaluation is based on manually labelled real-world data collected over 4 days of regular work life of the mobile workforce. The collected data yields 83 tasks in total involving 8 different orderlies from a major university hospital with a building area of 160, 000 m2. The results show that the proposed methods can distinguish accurately between the four most common task phases present in the orderlies' work routines, achieving Fi-Scores of 89.2%.

The increasing prevalence of positioning and tracking systems has helped simplify tracking large amounts of, e.g., People moving through buildings or cars traveling on roads, over long periods of time. However, technical limitations of positioning algorithms and traditional sensing infrastructures are likely, especially indoors, to induce errors and biases in the resulting data. In particular, the resulting motion trajectories often do not conform perfectly to the underlying route network. As a consequence, analyses of trajectory sets are impeded by these phenomena, as it becomes hard to identify which route was taken in a particular travel instance or whether two travel instances followed the same route. In this paper, we present a bootstrapping approach and several algorithms to mitigate error biases and related phenomena, focusing on indoor scenarios. In particular, we are able to estimate and iteratively refine an underlying route network from a set of motion trajectories. Secondly, we represent sub trajectories, i.e., Movements on individual elements of the route network, by their median sub trajectory. The resulting aggregated and cleaned-up data set facilitates using further, domain-specific analysis tools. Additionally, it allows to predict the locally occurring expected positioning error biases. This in turn allows improved positioning, e.g., For real-time navigation assistance scenarios. We evaluate the proposed methods using trajectory data from employees at a large hospital complex. In particular, we show that we can reconstruct the hospital's route network accurately, and that we can furthermore extract median sub trajectories for almost all individual corridors. Finally, we illustrate that median trajectories deliver useful deviation maps to learn, and correct for, the expected local biases in positioning.

The widespread presence of motion sensors on users' personal mobile devices has spawned a growing research interest in human activity recognition (HAR). However, when deployed at a large-scale, e.g., on multiple devices, the performance of a HAR system is often significantly lower than in reported research results. This is due to variations in training and test device hardware and their operating system characteristics among others. In this paper, we systematically investigate sensor-, device- and workload-specific heterogeneities using 36 smartphones and smartwatches, consisting of 13 different device models from four manufacturers. Furthermore, we conduct experiments with nine users and investigate popular feature representation and classification techniques in HAR research. Our results indicate that on-device sensor and sensor handling heterogeneities impair HAR performances significantly. Moreover, the impairments vary significantly across devices and depends on the type of recognition technique used. We systematically evaluate the effect of mobile sensing heterogeneities on HAR and propose a novel clustering-based mitigation technique suitable for large-scale deployment of HAR, where heterogeneity of devices and their usage scenarios are intrinsic.

Many mobile location-aware applications require the sampling of trajectory data accurately over an extended period of time. However, continuous trajectory tracking poses new challenges to the overall battery life of the device, and thus novel energy-efficient sensor management strategies are necessary for improving the lifetime of such applications. Additionally, such sensor management strategies are required to provide a high and application-adjustable level of robustness regardless of the user’s transportation mode. In this article, we extend and further analyze the sensor management strategies of the EnTracked$_{T}$system that intelligently determines when to sample different on-device sensors (e.g., accelerometer, compass and GPS) for trajectory tracking. Specifically, we propose the concept of situational bounding to improve and parameterize the robustness of sensor management strategies for trajectory tracking. We demonstrate the effectiveness of our proposed approach by performing a series of emulation experiments on real world data sets collected from different modes of transportation (including walking, running, biking and commuting by car) on mobile devices from two different platforms. Thorough experimental analyses indicate that our system can save significant amounts of battery power compared to the state-of-the-art position tracking systems, while simultaneously maintaining robustness and accuracy bounds as required by diverse location-aware applications.

Accurate information about how people commonly travel in a given large-scale building environment and which routes they take for given start and destination points is essential for applications such as indoor navigation, route prediction, and mobile work planning and logistics. In this paper, we propose methods for detecting commonly used routes by robust aggregation, clustering, and merging of indoor position traces. The developed methods overcome three specific challenges for detecting commonly used routes in an indoor setting based on position data: i) a high ratio between path-density and positioning-accuracy, ii) a flat path hierarchy, and iii) providing cost-effective scalability. Through an evaluation based on data collected by staff members at a hospital covering more than 10 hectare over three floors, we show that the proposed methods detect routes that are representative of the commonly used routes between locations. These methods are sufficiently efficient to provide common routes based on real-time data from thousands of devices simultaneously. Furthermore, we show that the methods operate robustly even on basis of noisy and coarse-grained position estimates as provided by large-scale deployable indoor Wi-Fi positioning systems, and with no prior information on building layout.

Existing methods for transportation mode detection (TMD) using mobile sensing make it generally possible to distinguish between walking, cycling, and motorized transport. However, our means of transport evolve and we develop radically new ways of transporting ourselves, thus new TMD sub-classification methods are needed to distinguish these new transport forms. As we transition from fossil-fueled cars to electric vehicles, switch to bikes with electric motors, ride in hybrid buses, or do city sightseeing on Segways, new challenges arise in distinguishing these from a mobile sensing perspective. Distinguishing electric vehicles (EVs) from fossil-fueled vehicles (FFVs) is a challenge, where traditional methods based on features such as GPS speed, or statistics on raw accelerometer data, are insufficient. In this paper, we present methods for distinguishing EVs from FFVs using smartphones with built-in inertial sensors, by reliably identifying idle-engine motor vibrations through features built on frequency analysis. We provide an extensive analysis of the challenges involved in making the EV/FFV distinction, as well as practical tools based on the methods. This includes analyzing the measurable similarities and differences between EVs and FFVs, and developing methods of reliably separating them. The presented tools implement the methods as classifiers built using machine learning. The analysis of our experiments shows that we can achieve an accuracy of 89-95% distinguishing EVs from FFVs, even with on-body phones.

The ability to accurately estimate indoor travel times is crucial for enabling improvements within application areas such as indoor navigation, logistics for mobile workers, and facility management. In this paper, we study the challenges inherent in indoor travel time estimation, and we propose th

The optimization of logistics in large building complexes with many resources, such as hospitals, require realistic facility management and planning. Current planning practices rely foremost on manual observations or coarse unverified assumptions and therefore do not properly scale or provide realistic data to inform facility planning. In this paper, we propose analysis methods to extract knowledge from large sets of network collected WiFi traces to better inform facility management and planning in large building complexes. The analysis methods, which build on a rich set of temporal and spatial features, include methods for noise removal, e.g., labeling of beyond building-perimeter devices, and methods for quantification of area densities and flows, e.g., building enter and exit events, and for classifying the behavior of people, e.g., into user roles such as visitor, hospitalized or employee. Spatio-temporal visualization tools built on top of these methods enable planners to inspect and explore extracted information to inform facility-planning activities. To evaluate the methods, we present results for a large hospital complex covering more than 10 hectares. The evaluation is based on WiFi traces collected in the hospital's WiFi infrastructure over two weeks observing around 18000 different devices recording more than a billion individual WiFi measurements. For the presented analysis methods we present quantitative performance results, e.g., demonstrating over 95% accuracy for correct noise removal of beyond building perimeter devices. We furthermore present detailed statistics from our analysis regarding people's presence, movement and roles, and example types of visualizations that both highlight their potential as inspection tools for planners and provide interesting insights into the test-bed hospital.

The vast availability of mobile phones with built-in movement and location sensors enables the collection of detailed information about human movement even indoors. As mobility is a key element of many processes and activities, an interesting class of information to extract is movement patterns that quantify how humans move, interact and group. In this paper we propose methods for detecting two common pedestrian movement patterns, namely individual following relations and group leadership. The proposed methods for identifying following patterns employ machine learning on features derived using similarity analysis on time-lagged sequences of WiFi measurements containing either raw signal strength values or derived locations. To detect leadership we combine the individual following relations into directed graphs and detect leadership within groups by graph link analysis. Methods for detecting these movement patterns open up new possibilities in–amongst others–computational social science, reality mining, marketing research and location-based gaming. We provide evaluation results that show error rates down to 7%, improving over state-of-the-art methods with up to eleven percentage points for following patterns and up to twenty percentage points for leadership patterns. Furthermore, we provide an analysis of the computational efficiency of the proposed methods and present visualizations for the analysis of detected patterns. Our methods are, contrary to state of the art, also applicable in challenging indoor environments, e.g., multi-story buildings. This implies that even quite small samples allow us to detect information such as how events and campaigns in multi-story shopping malls may trigger following in small groups, or which group members typically take the lead when triggered by e.g. commercials, or how rescue or police forces act during training exercises.

The vast availability of mobile phones with built-in movement and location sensors enable the collection of detailed information about human movement even indoors. As mobility is a key element of many processes and activities, an interesting class of information to extract is movement patterns that quantify how humans move, interact and group. In this paper we propose methods for detecting two common pedestrian movement patterns, namely individual following relations and group leadership. The proposed methods for identifying following patterns employ machine learning on features derived using similarity analysis on time lagged sequences of WiFi measurements containing either raw signal strength values or derived locations. To detect leadership we combine the individual following relations into directed graphs and detect leadership within groups by graph link analysis. Methods for detecting these movement patterns open up new possibilities in — amongst others — computational social science, reality mining, marketing research and location-based gaming. We provide evaluation results that show error rates down to 7%, improving over state of the art methods with up to eleven percentage points for following patterns and up to twenty percentage points for leadership patterns. Our method is, contrary to state of the art, also applicable in challenging indoor environments, e.g., multi-story buildings. This implies that even quite small samples allow us to detect information such as how events and campaigns in multistory shopping malls may trigger following in small groups, or which group members typically take the lead when triggered by e.g. commercials, or how rescue or police forces act during training exercises.

No abstract available.

Today, power consumption is a main limitation for mobile phones. To minimize the power consumption of popular and traditionally power-hungry location-based services requires knowledge of how individual phone features consume power, so that those features can be utilized intelligently for optimal power savings while at the same time maintaining good quality of service. This paper proposes an unsupervised API-level method for power profiling mobile phones based on genetic algorithms. The method enables accurate profiling of the power consumption of devices and thereby provides the information needed by methods that aim to minimize the power consumption of location-based and other services.

Emergent location-aware applications often require tracking trajectories of mobile devices over a long period of time. To be useful, the tracking has to be energy-efficient to avoid having a major impact on the battery life of the mobile device. Furthermore, when trajectory information needs to be sent to a remote server, on-device simplification of the trajectories is needed to reduce the amount of data transmission. While there has recently been a lot of work on energy-efficient position tracking, the energy-efficient tracking of trajectories has not been addressed in previous work. In this paper we propose a novel on-device sensor management strategy and a set of trajectory updating protocols which intelligently determine when to sample different sensors (accelerometer, compass and GPS) and when data should be simplified and sent to a remote server. The system is configurable with regards to accuracy requirements and provides a unified framework for both position and trajectory tracking. We demonstrate the effectiveness of our approach by emulation experiments on real world data sets collected from different modes of transportation (walking, running, biking and commuting by car) as well as by validating with a real-world deployment. The results demonstrate that our approach is able to provide considerable savings in the battery consumption compared to a state-of-the-art position tracking system while at the same time maintaining the accuracy of the resulting trajectory, i.e., support of specific accuracy requirements and different types of applications can be ensured.

We describe space-efficient algorithms for solving problems related to finding maxima among points in two and three dimensions. Our algorithms run in optimal $\mathcal{O}(n\log n)$ time and occupy only constant extra space in addition to the space needed for representing the input.