Miguel A. Labrador

A Survey on Human Activity Recognition using Wearable Sensors

Providing accurate and opportune information on peoples activities and behaviors is one of the most important tasks in pervasive computing. Innumerable applications can be visualized, for instance, in medical, security, entertainment, and tactical scenarios. Despite human activity recognition (HAR) being an active field for more than a decade, there are still key aspects that, if addressed, would constitute a significant turn in the way people interact with mobile devices. This paper surveys the state of the art in HAR based on wearable sensors. A general architecture is first presented along with a description of the main components of any HAR system. We also propose a two-level taxonomy in accordance to the learning approach (either supervised or semi-supervised) and the response time (either offline or online). Then, the principal issues and challenges are discussed, as well as the main solutions to each one of them. Twenty eight systems are qualitatively evaluated in terms of recognition performance, energy consumption, obtrusiveness, and flexibility, among others. Finally, we present some open problems and ideas that, due to their high relevance, should be addressed in future research.

A location-based incentive mechanism for participatory sensing systems with budget constraints

Participatory sensing (PS) systems rely on the willingness of mobile users to participate in the collection and reporting of data using a variety of sensors either embedded or integrated in their cellular phones. However, this new data collection paradigm has not been very successful yet mainly because of the lack of incentives for participation. Although several incentive schemes have been proposed to encourage user participation, none has used location information and imposed budget and coverage constraints, which will make the scheme more realistic and efficient. We propose a recurrent reverse auction incentive mechanism with a greedy algorithm that selects a representative subset of the users according to their location given a fixed budget. Compared to existing mechanisms, our incentive scheme improves the area covered by more than 60 percent acquiring a more representative set of samples after every round while maintaining the same number of active users in the system and spending the same budget.

Centinela: A human activity recognition system based on acceleration and vital sign data

This paper presents Centinela, a system that combines acceleration data with vital signs to achieve highly accurate activity recognition. Centinela recognizes five activities: walking, running, sitting, ascending, and descending. The system includes a portable and unobtrusive real-time data collection platform, which only requires a single sensing device and a mobile phone. To extract features, both statistical and structural detectors are applied, and two new features are proposed to discriminate among activities during periods of vital sign stabilization. After evaluating eight different classifiers and three different time window sizes, our results show that Centinela achieves up to 95.7% overall accuracy, which is higher than current approaches under similar conditions. Our results also indicate that vital signs are useful to discriminate between certain activities. Indeed, Centinela achieves 100% accuracy for activities such as running and sitting, and slightly improves the classification accuracy for ascending compared to the cases that utilize acceleration data only.

Survey on fall detection and fall prevention using wearable and external sensors.

According to nihseniorhealth.gov (a website for older adults), falling represents a great threat as people get older, and providing mechanisms to detect and prevent falls is critical to improve peoples lives. Over 1.6 million U.S. adults are treated for fall-related injuries in emergency rooms every year suffering fractures, loss of independence, and even death. It is clear then, that this problem must be addressed in a prompt manner, and the use of pervasive computing plays a key role to achieve this. Fall detection (FD) and fall prevention (FP) are research areas that have been active for over a decade, and they both strive for improving peoples lives through the use of pervasive computing. This paper surveys the state of the art in FD and FP systems, including qualitative comparisons among various studies. It aims to serve as a point of reference for future research on the mentioned systems. A general description of FD and FP systems is provided, including the different types of sensors used in both approaches. Challenges and current solutions are presented and described in great detail. A 3-level taxonomy associated with the risk factors of a fall is proposed. Finally, cutting edge FD and FP systems are thoroughly reviewed and qualitatively compared, in terms of design issues and other parameters.

P-Sense: A participatory sensing system for air pollution monitoring and control

This article presents P-Sense, a participatory sensing application for air pollution monitoring and control. The paper describes in detail the system architecture and individual components of a successfully implemented application. In addition, the paper points out several other research-oriented problems that need to be addressed before these applications can be effectively implemented in practice, in a large-scale deployment. Security, privacy, data visualization and validation, and incentives are part of our work-in-progress activities

Packet dropping policies for ATM and IP networks

Selective packet dropping policies have been used to reduce congestion and transmission of traffic that would inevitably be retransmitted. For data applications using best-effort services, packet dropping policies (PDPs) are congestion management mechanisms implemented at each intermediate node that decide, reactively or proactively, to drop packets to reduce congestion and free up precious buffer space. While the primary goal of PDPs is to avoid or combat congestion, the individual PDP designs can significantly affect application throughput, network utilization, performance fairness, and synchronization problems with multiple transmission control protocol (TCP) connections. Scalability and simplicity are also important design issues. This article surveys the most important selective packet dropping policies that have been designed for best-effort traffic in ATM and IP networks, providing a comprehensive comparison between the different mechanisms.

On the applicability of available bandwidth estimation techniques and tools

Available Bandwidth Estimation Techniques and Tools (ABETTs) have recently been envisioned as a supporting mechanism in areas such as compliance of service level agreements, network management, traffic engineering and real-time resource provisioning, flow and congestion control, construction of overlay networks, fast detection of failures and network attacks, and admission control. However, it is unknown whether current ABETTs can run efficiently in any type of network, under different network conditions, and whether they can provide available bandwidth estimates at the timescales needed by these applications. This article includes a performance evaluation of Pathload, Pathchirp, Spruce, IGI, and Abing in a low cost and flexible test bed. The evaluation includes scenarios and conditions not evaluated before, such as varying the packet loss rate and the propagation delays of the links, the amount of cross-traffic, the capacity of the links, and the cross-traffic packet size. The results demonstrate that ABETTs are far from being ready to be applied in all these applications and scenarios. In addition, the article clearly indicates those aspects that need further research and which ABETTs are the best candidates for specific applications and environments.

G-Sense: a scalable architecture for global sensing and monitoring

The pervasiveness of cellular phones combined with Internet connectivity, GPS embedded chips, location information, and integrated sensors provide an excellent platform to collect data about the individual and its surrounding environment. As a result, new applications have recently appeared to address large-scale societal problems as well as improve the quality of life of the individual. However, these new applications, recently called location-based services, participatory sensing, and human-centric sensing, bring many new challenges, one of them being the management of the huge amount of traffic (data) they generate. This article presents G-Sense, for Global-Sense, an architecture that integrates mobile and static wireless sensor networks in support of location-based services, participatory sensing, and human-centric sensing applications. G-Sense includes specific mechanisms to control the amount of data generated by these applications while meeting the application requirements. Furthermore, it creates a network of servers organized in a peer-to-peer architecture to address scalability and reliability issues. An example prototype application is presented along with some basic results and open research issues.

A3: A Topology Construction Algorithm for Wireless Sensor Networks

Topology control is a well-known strategy to save energy and extend the lifetime of wireless sensor networks. This paper introduces the A3 (a tree) algorithm, a simple, distributed, and energy-efficient topology construction mechanism that finds a sub-optimal Connected Dominating Set (CDS) to turn unnecessary nodes off while keeping the network connected and providing complete communication coverage. A3 utilizes a weighted distance-energy-based metric that permits the network operator to trade off the lengths of the branches (distance) for the robustness and durability of the tree (energy). Comparisons with other well-known topology construction mechanisms show the superiority of the proposed scheme in terms of the number of active nodes and energy efficiency. Simulation experiments show that to achieve complete communication coverage, A3 needs only 6% and 41% of the nodes active in dense and sparse scenarios, versus 8% and 43% and 5% and 43% of the EECDS and CDS- Rule-K algorithms, respectively. More importantly, the proposed protocol presents a low linearly bounded worst-case amount of messages per node that limits the overhead and the energy usage compared to a non-linear increase of the EECDS and CDS-Rule- K algorithms.

A mobile platform for real-time human activity recognition

Context-aware applications have been the focus of extensive research yet their implementation in mobile devices usually becomes challenging due to restrictions in regards to processing power and energy. In this paper, we propose a mobile platform to provide real-time human activity recognition. Our system features (1) an efficient library, MECLA, for the mobile evaluation of classification algorithms; and (2) a mobile application for real-time human activity recognition running within a Body Area Network. The evaluation indicates that the system can be implemented in real scenarios meeting accuracy, response time, and energy consumption requirements.