The XAC project focuses on developing shared WSNs that can be used by multiple context-aware systems. Our research is categorized into two parts: self-adaptive WSN middleware and model-driven development for WSN applications.

1. Self-adaptive WSN middleware
   A sensor node in a WSN has limited resources, e.g. CPU, memory, bandwidth, and battery. Still, a middleware for a shared WSN has to provide QoS for each application (e.g. accuracy of observation) under such resource constraints. For this reason, we have proposed self-adaptive methods to reduce redundant resource consumptions. For instance, we have presented a configurable sensor model that enables to adjust the quality of the observation. Also, we have proposed a sensor selection method to minimize the number of nodes participating in the observation along with a sensor allocation method to guarantee fairness among multiple applications. Since WSNs intrinsically have a dynamic nature, these methods are required to be self-adaptive.
2. Model-driven development for WSN application
   When WSNs are shared, the WSN application developer can reduce the development costs of applications by ignoring system-specific aspects of one WSN. Therefore, we have proposed a model-driven WSN application development method that enables the developer to describe the core logic in a WSN-independent model. Subsequently, this model is transformed automatically into executable code by using a model transformation engine and rules.

Research Topics

Configurable Sensor Model for Target Tracking in Wireless Sensor Networks

In shared wireless sensor networks a vast amount of data is transferred between the nodes. In order to improve the scalability, the data transmissions caused by each application should be reduced to a minimum while upkeeping the required resolution. Existing works proposed a binary sensor model, which maps the sensory data to binary values to reduce the necessary data transmissions by sacrificing resolution at the same time. However, this binary sensor model cannot adjust the resolution adequately to the required resolution. Therefore, we proposed a N-ary sensor model by generalizing the binary sensor model. The N-ary sensor model maps sensory data to N-ary values. This way, the model can adjust the resolution of each sensor node by configuring the arity N. In addition, we proposed an arity configuration algorithm to assign an appropriate arity to each node in order to further reduce needless data transmissions while preserving the required resolution.

Geographic-based Adaptive Sensor Node Selection

In a sensor network, limiting the number of sensors used for observation purposes is an effective way to reduce the energy consumption of each sensor. In order to limit the number of sensors without sacrificing observation accuracy, an appropriate sensor combination must be selected by evaluating the observation effectiveness of various combinations. However, the computational workload for evaluating all the sensor combinations is quite large. In region-based sensor selection, a combination of sensors that are near to the observation target is selected. We define a parameter related to the optimal size of a region around an observation target by making a trade-off between accuracy and computational workload.

Sensor Node Assignment for Multiple Tasks

Wireless Sensor Networks are applied as a platform for ubiquitous computing. Since WSNs are deployed in steadily growing numbers and are used for many purposes, a huge amount of different tasks with a variety of priorities/requirements are deployed into such networks. In general, tasks are executed by assigning them to a concrete set of sensors. However, it remains unclear how to assign the sensors to the tasks such that both, the requirements of the tasks and the constraints on the network, are satisfied. In this research work, we first formulate this problem as an optimization problem by defining several constraints on the WSN. This way, we can thoroughly evaluate several algorithms by giving bounds for the problem. Furthermore, by extracting the features of the formulated problem, we can propose a new algorithm that has a better performance. In short, the goal of this research is that in the future the dwarf shown in the picture, who is responsible for the sensor allocation, makes smarter decisions.

Self-Corrective Fault Tolerant Wireless Sensor Networks

Ultimately, the goal of WSN is to provide accurate data about monitored phenomena efficiently over the maximum possible period. General characteristics of wireless sensor networks and the direct exposure to the environment cause frequent faults. Accumulation of these faults can lead to the progressive decrease of the reliability and accuracy of sensor readings. This leads to the shortening of effective lifetime, defined as the time of operation in which network reliably provides accurate data. An automated framework for fault tolerance would enable the network to be aware of faults and adapt and correct its behavior accordingly. Fault detection and classification based on continuity and frequency of occurrence is the starting point for discovery of observable and learnable patterns in faulty readings. Models of faults learned in this phase are used to correct readings from faulty sensors. In this way, network learns from observing its own behavior and learns how to maximize the utility of each sensor node.

Model-Driven Development for Rapid Prototyping and Optimization of Wireless Sensor Network Applications

In order to develop Wireless Sensor Network (WSN) applications, it is necessary to develop prototypes in a low-cost way and to optimize the performance of the applications. Existing development approaches enable to create low-cost prototypes because they reduce the components described by the developers. However, there is a trade-off between the costs of developing a prototype and the description capabilities needed to optimize the performance of the application. We propose a Model-Driven Development (MDD) process to enable a low-cost prototyping and a detailed optimization of the performance. To enable such a development process, we define modeling languages which describe an application at three abstraction levels, and associated transformation rules adding information that appears in the output model for the first time. Using our process for prototyping the developer describes a model by using the modeling language at the highest abstraction level and automatically obtains an executable model by using the transformation rules. In addition, in the optimizing process the developer can automatically obtain the models at a more concrete abstraction level than the prototype by using transformation rules and modify them in greater detail by using each modeling language.

Members

• Kenji Tei (Assistant Professor)
• Yoshiyuki Nakamura (D4)
• Susumu Toriumi(D3)
• Valentina Baljak (D3)
• Ryo Shimizu (D1)

Contact Information

Kenji Tei：

Research Results

Selected Publications

1. Ryo Shimizu, Kenji Tei, Yoshiaki Fukazawa and Shinichi Honiden: "Case Studies on the Development of Wireless Sensor Network Applications using Multiple Abstraction Levels", In Proceedings of the 3rd International Workshop on Software Engineering for Sensor Network Applications (SESENA 2012), in conjunction with International Conference on Software Engineering (ICSE) June 2-9, 2012. (to appear)

2. Ehsan Warriach, Kenji Tei, Tuan Anh Nguyen, Marco Aiello: "Fault Detection in Wireless Sensor Networks: a Hybrid Approach", Poster session of the 11th ACM/IEEE Conference on Information Processing in Sensor Networks (IPSN'12), April 2012.

3. Ryo Shimizu, Kenji Tei, Yoshiaki Fukazawa, Shinichi Honiden: "Meta-Models for Wireless Sensor Network Applications: Data, Group, and Node Views", Technical Report GRACE-TR-2012-01, GRACE Center, National Institute of Informatics, Frbruary 2012. 9 pages.

4. Susumu Toriumi, Shinichi Honiden: "Assignment of Sensors for Multiple Tasks Using Path Information", 9th IEEE/IFIP International Conference on Embedded and Ubiquitous Computing (EUC-2011), 120-127, October 2011

5. Themistoklis Bourdenas, Kenji Tei, Shinichi Honiden and Morris Sloman: "Autonomic Role and Mission Allocation Framework for Wireless Sensor Networks", Fifth IEEE International Conference on Self-Adaptive and Self-Organizing Systems (SASO'11), October 3 - October 7, 2011.

6. Ryo Shimizu, Kenji Tei, Yoshiaki Fukazawa, Shinichi Honiden: "Model Driven Development for Rapid Prototyping and Optimization of Wireless Sensor Network Applications", In Proceedings of the 2nd International Workshop on Software Engineering for Sensor Network Applications (SESENA 2011), in conjunction with ACM/IEEE Intl. Conference on Software Engineering (ICSE) May 21-28, 2011

7. Rey Abe, Shinichi Honiden: "Suppressing Redundancy in Wireless Sensor Network Traffic", In Proceedings of the 6th IEEE International Conference on Distributed Computing in Sensor Systems, Santa Barbara (DCOSS' 10), California, USA. Springer-Verlag, June 21-23, 2010.

8. Valentina Baljak, Shinichi Honiden: "Discovery of Configurations for Indoor Wireless Sensor Networks Through Use of Simulation in Virtual Worlds", In Proceedings of the Fourth International Conference on Sensor Technologies and Applications (SENSORCOMM 2010). July 18-25, 2010.

9. Rey Abe, Shinichi Honiden: "Adaptive Geographic Routing in Wireless Sensor Networks", In Proceedings of the 13th ACM International Conference on Modeling, Analysis and Simulation of Wireless and Mobile Systems (MSWiM 2010). October 17-21, 2010.

10. Kenji Tei, Shunichiro Suenaga, Yoshiyuki Nakamura, Yuichi Sei, Hikotoshi Nakazato, Yoichi Kaneki, Nobukazu Yoshioka, Yoshiaki Fukazawa, Shinichi Honiden: "XAC Project: Towards a Middleware for Open Wireless Sensor Networks", chapter in book "Designing Solutions-Based Ubiquitous and Pervasive Computing: New Issues and Trends", edited by Francisco Milton Mendes Neto and Pedro Fernandes Ribeiro Neto. published by Information Science Publishing, March 2010.

Developed softwares

Wireless sensor network middleware for SunSPOT：XACMiddleware