This paper presents a major step in the development and validation of a systematic prototype-based methodology for designing multilayer feedforward neural networks to model nonlinearities common in engineering mechanics. The applications of this work include (but are not limited to) system identification of nonlinear dynamic systems and neural-network-based damage detection. In this and previous studies (Pei, J. S., 2001, “Parametric and Nonparametric Identification of Nonlinear Systems,” Ph.D. thesis, Columbia University; Pei, J. S., and Smyth, A. W., 2006, “A New Approach to Design Multilayer Feedforward Neural Network Architecture in Modeling Nonlinear Restoring Forces. Part I: Formulation,” J. Eng. Mech., 132(12), pp. 1290–1300; Pei, J. S., and Smyth, A. W., 2006, “A New Approach to Design Multilayer Feedforward Neural Network Architecture in Modeling Nonlinear Restoring Forces. Part II: Applications,” J. Eng. Mech., 132(12), pp. 1301–1312; Pei, J. S., Wright, J. P., and Smyth, A. W., 2005, “Mapping Polynomial Fitting Into Feedforward Neural Networks for Modeling Nonlinear Dynamic Systems and Beyond,” Comput. Methods Appl. Mech. Eng., 194(42–44), pp. 4481–4505), the authors do not presume to provide a universal method to approximate any arbitrary function. Rather the focus is given to the development of a procedure which will consistently lead to successful approximations of nonlinear functions within the specified field. This is done by examining the dominant features of the function to be approximated and exploiting the strength of the sigmoidal basis function. As a result, a greater efficiency and understanding of both neural network architecture (e.g., the number of hidden nodes) as well as weight and bias values is achieved. Through the use of illuminating mathematical insights and a large number of training examples, this study demonstrates the simplicity, power, and versatility of the proposed prototype-based initialization methodology. A clear procedure for initializing neural networks to model various nonlinear functions commonly seen in engineering mechanics is provided. The proposed methodology is compared with the widely used Nguyen–Widrow initialization to demonstrate its robustness and efficiency in the specified applications. Future work is also identified.

No abstract

No abstract

Pavement maintenance is vital for travel safety, thus detecting dangerous road conditions in a real-time fashion is desirable. Using an off-the-shelf wireless sensor network to detect such conditions at a low cost poses many challenges. In order to meet these challenges, a Mica2 Mote sensor network is adopted in this study to process and transmit data collected from three external analog sensors. Consequentially, several hardware and software interfaces are developed to complete a pavement monitoring system that uses temperature and moisture presence to detect hazardous road conditions. Surge Time Synchronization is explored in this specific application to enable the wireless sensor network to operate in a low power consumption mode. A fairly simplistic pattern classification algorithm is embedded into the motes to create the smart wireless sensing application. A series of outdoor tests are conducted in this study paying special attention to the survivability of fragile analog sensors in harsh roadway conditions. In this regard, a novel solution called the ''Sensor-Road Button''(SRB) is developed and validated experimentally. This is one of several exercises made in this study to enable the application of sensor technologies in intelligent transportation systems (ITS). The size of the wireless sensor network in this study is relatively small, utilizing a total of five motes in order to fully exploit the transmitting range of each mote. Long testing periods (i.e., uninterrupted 12-hour time frames for each period of data collection) add an additional advantage, allowing for the evaluation of the selected wireless sensor network for long-term monitoring using the low power consumption mode under Surge Time Synchronization. Many performance metrics of the adopted small-size, large-interval Mica2 Motes wireless sensor network are revealed in this study through a series of data processing efforts. Results are presented to examine (i) inter-node connectivity and transmitting range, (ii) battery life, (iii) the length of the initial network connection time as affected by methods of setting up tests under practical conditions, (iv) error rate and analysis of different error types (showing the importance of the subsequent data cleansing step), and (v) other network routing properties including the parent time histories for each mote. The results and analysis form a database for future efforts to better understand, appreciate, and improve the performance of Mica2 Motes. This study will thus benefit robust real-world implementation of off-the-shelf sensor network products such as Mica2 Motes in terms of hardware development and data processing.

No abstract