Kyungseo Park

Abnormal human behavioral pattern detection in assisted living environments

In recent years, there is a growing interest about assisted living environments especially for the elderly who live alone, due to the increasing number of aged people. In order for them to live safe and healthy, we need to detect abnormal behavior that may cause severe and emergent situations for the elderly. In this work, we suggest a method that detects abnormal behavior using wireless sensor networks. We model an episode that is a series of events, which includes spatial and temporal information about the subject being monitored. We define a similarity scoring function that compares two episodes taking into consideration temporal aspects. We propose a way to determine a threshold to divide episodes into two groups that reduces wrong classification. Weights on individual functions that consist the similarity function are determined experimentally so that they can produce the good results in terms of area under curve in receiver operating characteristic (ROC) curve.

Human behavioral detection and data cleaning in assisted living environment using wireless sensor networks

Due to the increasing number of the elderly, more and more people need to have additional health care such as medical or environmental monitoring information at home or nursing facility. Most elderly people are likely to have a sudden behavioral changes due to their aging or existing health problems. Therefore, it is necessary to have an autonomous system that can monitor them in order to prevent emergent situation in advance. In this paper, we present a wireless sensor network system that can recognize human behavioral patterns of the elderly who lives alone. We model episodes that are series of events for a person who lives in an one-bedroom apartment. We propose data cleaning techniques in both sensor and base station sides for the erroneous environment of wireless sensor networks. Based on these techniques, we try to extract discrete events as close as possible to effective events. We introduce non-real time analysis to recognize human behavioral patterns on the centralized system, which can be further extended to a real-time analysis. We also adopt an existing search technique to apply it to detect similar or abnormal behavior. We experiment the proposed system by gathering behavioral pattern data from the miniature one-bedroom apartment that is equipped with SunSPOTs in our HERACLEIA Laboratory. We look up the resulting episodes from our experiment in the dictionary that is a set of predetermined episodes using the suggested algorithm.

Event-based experiments in an assistive environment using wireless sensor networks and voice recognition

As the population is aging, more and more people require additional health care, either at home, in the work place or in a nursing facility. Now, a need exists for health monitoring outside of hospital conditions. These new conditions make this technology of interest for developing health care monitoring systems that can be deployed in many different environments, including the home. Other systems in development employ a wide range of different sensors, including cameras, and recording the information for processing. These systems all involve using an apartment environment seeded with sensors for detecting human behavior and activities. While these systems are embedded in assistive environments, they do not have a comprehensive approach to describe events, or handle a general and rapid deployment into different configurations using wireless technology. In this paper, we are presenting our ongoing project of deploying sensors into an assistive environment. We currently are using SunSPOT sensor motes, where each one has been programmed for a specific role based on rules describing events. In addition, we are developing a voice recognition system for reaction to human input in the same environment. Our system can be rapidly deployed without requiring additional wiring or unwanted intrusion into the human patients life.

Energy Efficient Spatial Query Processing in Wireless Sensor Networks

Because a sensor network depends on limited battery power, energy saving is important to increase the sensor network lifespan. We propose semi-distributed spatial query indexing structure that disseminates a query into the network and retrieves data energy efficiently using a localized tree building algorithm. We also propose a sectioned tree index, which divides the network area into several squares and each square has a local index subtree organized within that square. Local trees are interconnected to form one big tree in the network. Local trees are also built based on any algorithm that is energy consumption aware at each sub-root node in a locally centralized way. We use an existing two dimensional indexing technique for energy efficient query dissemination. We show that our proposed scheme is energy efficient for query and data processing heuristically. Our proposed scheme, sectioned tree, is finally simulated in sparse and dense networks to show the energy saving for query and data processing in the sensor network.

Decision making in assistive environments using multimodal observations

An assistive environment is a smart domestic space based on pervasive computing to support the elderly and disabled. Unlike sensors, which can only provide passive monitoring, a robot can be an active element to improve the quality of life for the human. In this paper, we propose an active service of the robot in assistive environments, to help human in the case of emergency situation. It works on a hierarchical partially observable Markov decision process (POMDP). The multimodal observation series are used in the decision and evaluation process. An active robot is a kind of robot that can provide a preferable and necessary active service to the human. This is used in our emergency response system (ERS) to deal with the emergency situations, such as an older adult falls down or emergency diseases. The purpose of multimodal observations is to guarantee the precision of report for the emergency situations. Four observation sources are introduced in this paper: the vision recognition, the voice recognition, the physical input devices and the foreign systems. For each observation source, there are two observation series. Multiple information sources give the agent more opportunities to learn from the real world, so as to make more reasonable predictions, evaluations and decisions.

Architectures for streaming data processing in sensor networks

Summary form only given. Over the past few years, the development of technology has allowed new advances in sensor networks that monitor the physical world. The data streams produced by sensor networks have different characteristics from the data of traditional data processing, thus requiring new paradigms of data processing systems. Recently, a lot of research has been reported on data stream processing systems. However, there are many underlying assumptions about these systems that have not been explicitly specified. In this paper, our attempt is to provide a general model and architecture for data stream processing in sensor networks. This can serve as a reference architecture to better understand and categorize research in this area.

Query Classification and Storage Evaluation inWireless Sensor Networks

Recently, storage management has been a new focus of study in addition to traditional characteristics of wireless sensor networks. In this paper, five storage schemes that we categorize are evaluated in terms of query types that have different characteristics from one another. First, sensor network queries are classified by four criteria, and they are divided in more detail by the combination of elements in the criteria. Second, storage schemes are classified by considering where and how to store sensed data and what kind of routing protocols they use. Finally, storage schemes are evaluated for different query types in terms of metrics such as the number of transmissions, energy, delay, life span, and local storage capacity. Through the analysis and simulation, we show what kind of storage is suitable for which particular query characteristics.

Energy Balanced In-Network Aggregation Using Multiple Trees in Wireless Sensor Networks

Advances in wireless networks are expected to play an increasing role in systems that are aimed at collecting information. One of the main challenges in wireless sensor networks is that a sensor node has limited battery power. Therefore in order to increase the lifetime of sensor nodes, we need to reduce the amount of energy consumption. For reducing energy consumption in sensor networks, in-network aggregation is one of the proposed methods. However in-network aggregation does not keep the energy balance if some nodes are on the most frequently used paths in a network such as sink node. In order to consider more energy efficiency through load balancing, we propose a new in-network aggregation structure based on multiple trees, called MULT, for further extending the lifetime of in-network aggregation. Unlike existing in-network aggregation structures, which aim to reduce communication cost, the proposed MULT further provides energy balance. MULT has 3phases: first building the clusters, second connecting the clusters and third making multiple trees. MULT is based on creating node clusters using distance between nodes. In addition, a new clustering method, called HYC (HYbrid Cluster) is introduced for MULT structure. We compare the MULT with LEACH and EAD, which are popular in-network aggregation methods. MULT outperforms LEACH and EAD for energy load balance. KeywordsSensor Network, Energy Load Balance, In-Network aggregation

Effects of storage architecture on performance of sensor network queries

Storage architecture in sensor networks is increasingly emphasized as an important characteristic, in addition to more traditional characteristics like routing protocols and data dissemination techniques In this paper, we evaluate several types of storage in order to determine performance correlations between storage types and query types We first classify the various types of query and storage architectures for sensor networks We then evaluate storage architecture performance based on types of query The evaluation metrics we use are the number of transmissions, energy, and end-to-end delay Data delivery types and routing schemes have to also be considered since they are strongly related to the storage architecture Based on the performance evaluations, we show what kind of storage is suitable for particular query characteristics.

Reducing Energy Consumption through the Union of Disjoint Set Forests Algorithm in Sensor Networks

Recently, wireless sensor networks have improved for many applications aimed at collecting information. However wireless sensor networks have many challenges to be solved. One of the most critical problems is the energy restriction. Therefore in order to extend the lifetime of sensor nodes, we need to minimize the amount of energy consumption. In many cases, sensor networks use routing schemes based on the tree routing structure. But when we collect information from a restricted area within the sensor field using the tree routing structure, the information is often assembled by sensor nodes located on different tree branches. In this case unnecessary energy consumption happens in ancestor nodes located out of the target area. In this paper, we propose the Sensor Network Subtree Merge algorithm, called SNSM, which uses the union of disjoint set forest algorithm for preventing unnecessary energy consumption in ancestor nodes for routing. SNSM algorithm has 3-phases: first finding the disjoint set of the subtree in the sensor field; second connecting each disjoint subtree with the closest node; and third virtually disconnect the subtree connected to new tree branch from previous tree structure. In the simulation, we apply SNSM algorithm to a minimum spanning tree structure. Simulation results show that SNSM algorithm reduces the energy consumption. Especially, SNSM is more efficient as number of sensor nodes in a sensor field increases.