Elsa M. Macías

Mobile Sensing Systems

Rich-sensor smart phones have made possible the recent birth of the mobile sensing research area as part of ubiquitous sensing which integrates other areas such as wireless sensor networks and web sensing. There are several types of mobile sensing: individual, participatory, opportunistic, crowd, social, etc. The object of sensing can be people-centered or environment-centered. The sensing domain can be home, urban, vehicular… Currently there are barriers that limit the social acceptance of mobile sensing systems. Examples of social barriers are privacy concerns, restrictive laws in some countries and the absence of economic incentives that might encourage people to participate in a sensing campaign. Several technical barriers are phone energy savings and the variety of sensors and software for their management. Some existing surveys partially tackle the topic of mobile sensing systems. Published papers theoretically or partially solve the above barriers. We complete the above surveys with new works, review the barriers of mobile sensing systems and propose some ideas for efficiently implementing sensing, fusion, learning, security, privacy and energy saving for any type of mobile sensing system, and propose several realistic research challenges. The main objective is to reduce the learning curve in mobile sensing systems where the complexity is very high.

Ubiquitous Monitoring of Electrical Household Appliances

The number of appliances at home is increasing continuously, mainly because they make our lives easier. Currently, technology is integrated in all objects of our daily life. TCP/IP let us monitor our home in real time and check ubiquitously if something is happening at home. Bearing in mind this idea, we have developed a low-cost system, which can be used in any type of electrical household appliance that takes information from the appliance and posts the information to the Twitter Social network. Several sensors placed in the household appliances gather the sensed data and send them wired or wirelessly, depending on the case, using small and cheap devices to a gateway located in the home. This gateway takes decisions, based on the received data, and sends notifications to Twitter. We have developed a software application that takes the values and decides when to issue an alarm to the registered users (Twitter friends of our smart home). The performance of our system has been measured taking into account the home network (using IEEE 802.3u and IEEE 802.11g) and the data publishing in Twitter. As a result, we have generated an original product and service for any electrical household appliance, regardless of the model and manufacturer, that helps home users improve their quality of life. The paper also shows that there is no system with the same innovative features like the ones presented in this paper.

Gradient RSSI Filter and Predictor for Wireless Network Algorithms and Protocols

Wireless networks, and in particular Wireless Fidelity (WiFi) networks, present important problems of connectivity among wireless terminals. These problems lead to inefficiencies in the Application level. For localization algorithms in wireless networks, efficient handover protocols and multimedia delivery information, these problems must be solved. A figure used to control the connectivity of terminals to other terminals or to an Access Point (AP) is the well known Received Signal Strength Indicator (RSSI). Managing this figure it is possible to control the availability of a service in the wireless terminal. But the problem with RSSI is the fact that it exhibits high variability in space and time. Thus it is important to filter those values in order to efficiently manage the availability of the service in the mobile terminal. For that reason in this paper we present the experimental validation of a filter and a predictor of availability of a service in the wireless terminal based on the RSSI. We test the performance of the filter and the predictor that improves significantly other filters that are also presented in the paper. The significance of our work is that our filter allows simplifying a lot of network algorithms and protocols. As an example we present a simple protocol of handover in presence of multimedia information using our filter and predictor based on RSSI.

An Efficient Algorithm for Multiple Sclerosis Lesion Segmentation from Brain MRI

We propose a novel method for the segmentation of Multiple Sclerosis (MS) lesions in MRI. The method is based on a three-step approach: first a conventional k-NN classifier is applied to pre-classify gray matter (GM), white matter (WM), cerebro-spinal fluid (CSF) and MS lesions from a set of prototypes selected by an expert. Second, the classification of problematic patterns is resolved computing a fast distance transformation (DT) algorithm from the set of prototypes in the Euclidean space defined by the MRI dataset. Finally, a connected component filtering algorithm is used to remove lesion voxels not connected to the real lesions. This method uses distance information together with intensity information to improve the accuracy of lesion segmentation and, thus, it is specially useful when MS lesions have similar intensity values than other tissues. It is also well suited for interactive segmentations due to its efficiency. Results are shown on real MRI data as wall as on a standard database of synthetic images.

Architecture and protocol of a semantic system designed for video tagging with sensor data in mobile devices.

Current mobile phones come with several sensors and powerful video cameras. These video cameras can be used to capture good quality scenes, which can be complemented with the information gathered by the sensors also embedded in the phones. For example, the surroundings of a beach recorded by the camera of the mobile phone, jointly with the temperature of the site can let users know via the Internet if the weather is nice enough to swim. In this paper, we present a system that tags the video frames of the video recorded from mobile phones with the data collected by the embedded sensors. The tagged video is uploaded to a video server, which is placed on the Internet and is accessible by any user. The proposed system uses a semantic approach with the stored information in order to make easy and efficient video searches. Our experimental results show that it is possible to tag video frames in real time and send the tagged video to the server with very low packet delay variations. As far as we know there is not any other application developed as the one presented in this paper.

Programming Parallel Applications with LAMGAC in a LAN-WLAN Environment

Traditionally, the Local Area Network (LAN) has been used for parallel programming with PVM and MPI. The improvement of communications in Wireless Local Area Network (WLAN) achieving till 11 Mbps make them, according to some authors, candidates to be used as a resource for Grid Computing. Although CONDOR-MW tool can manage wireless communications using PVM (it is intended the use of MPI in future), in this paper we present a library based on LAM/MPI named LAMGAC in order to ease the programming of LAN-WLAN infrastructures. The novel of this research is that we can vary the parallel virtual machine in runtime, we generate algorithms in which computations and communications are efficiently overlapped and we include a web interface to offer our system as a Grid resource. At this moment we have measured the execution time of some algorithms and the functions of LAMGAC, obtaining interesting results.

A Hierarchical Communication Architecture for Oceanic Surveillance Applications

The interest in monitoring applications using underwater sensor networks has been growing in recent years. The severe communication restrictions imposed by underwater channels make that efficient monitoring be a challenging task. Though a lot of research has been conducted on underwater sensor networks, there are only few concrete applications to a real-world case study. In this work, hence, we propose a general three tier architecture leveraging low cost wireless technologies for acoustic communications between underwater sensors and standard technologies, Zigbee and Wireless Fidelity (WiFi), for water surface communications. We have selected a suitable Medium Access Control (MAC) layer, after making a comparison with some common MAC protocols. Thus the performance of the overall system in terms of Signals Discarding Rate (SDR), signalling delay at the surface gateway as well as the percentage of true detection have been evaluated by simulation, pointing out good results which give evidence in applicability’s favour.

Full Geo-localized Mobile Video in Android Mobile Telephones

The evolution of mobile telephones have produced smart devices that not only allows the mobile user to talk but also to use a lot of telematic services. High quality photos and videos are produced by smart mobile telephones. The Global Positioning System available in the Mobile telephones allows the user to tag their photos and videos. There are several photo and integral video tagging mobile software but there is not a mobile application that allows the mobile users to tag the full video frames. This full tagging process allows the mobile user to tag independent video frames in order to explode the photo-video properties of the integral video. In this paper we present a mobile application and a Server application that allow the mobile user to full tag the mobile videos and share them with other users (registered in the Server). We present some tradeoffs present in the design of the tagging process.

Solving Engineering Applications with LAMGAC over MPI-2

Traditionally, the Local Area Network (LAN) has been used for parallel programming. However, with recent advances in wireless communications, some authors think that the future of Wireless Local Area Networks (WLAN) depends on the demonstration of its application to real problems. In this paper we prove that wireless communications are suitable to perform parallel computing by presenting interesting experimental results for two engineering applications that follows the Master-Slave style programming with data dependencies among iterations. We consider a dynamic variation of slave processes as new portable nodes are added to the WLAN or detached from it.

Adaptive Estimation of WiFi RSSI and Its Impact Over Advanced Wireless Services

RSSI is a reference magnitude of radio signal strength received by a wireless terminal. Traditionally, it has been used to implement wireless services like scanning in handoff. But its high long term variability (volatility) makes its high precision estimation be impossible, complicating its applicability. Moreover, every wireless terminal has its own way to provide RSSI. Recently a high impact paper has formally shown that RSSI of laboratory radio signal in WiFi always reverts to its mean. That is, its long term stability can be estimated. We present a new RSSI estimation model that improves other recently published methods; and shows that it can be used, in contrast with those other techniques, to a wide range of current wireless terminals and real world scenarios. We also show its applicability to indoor localization service, by observing how it can be used in a recent high impact paper and improving another paper also published in a high impact journal. Our method can be used in SDN.