Peng Zhuang

SMART: Simultaneous indoor localization and map construction using smartphones

In recent years, we have witnessed a gold rush of location based services enabled by location awareness and persistent Internet connections of modern smartphones. However, the poor performance of indoor localization remains an obstacle to providing these services effectively inside buildings. Obtaining a comprehensive database of points of interest and high quality indoor maps is another obstacle because extensive human involvement is usually required for such tasks. In this paper, we propose a novel system SMART (Simultaneous Map Acquisition and Repeated Tracking) to address both problems. By tracking a subject based on sensor inputs and radio signals and detecting environmental fingerprints, SMART achieves simultaneous indoor localization and environment map construction using smartphones based on radio signals from surrounding WiFi access points and their own measured motion dynamics. Simulation results show that SMART outperforms a dead reckoning method by approximately 9 times in localization accuracy and constructs environment maps of 89% accuracy on average. SMART is robust against sensing errors and can automatically adapt to environment changes.

Distributed Faulty Sensor Detection

Frequently appeared sensor faults greatly reduce the usability and reliability of sensor networks. Distributed automatous faulty sensor detection is critical for self-managed and sustainable sensor networks. A number of detection methods have been proposed for specific fault types. In this paper, we propose a general approach for detecting arbitrary types of faults. The approach includes a new general measurement mutual divergence for evaluating detecting results in the absence of the ground truth, and a distributed collective detection method that produces probabilistic decision results. For comparison purposes, we also introduce two detection methods in both distributed and centralized manners. We show that mutual divergence correctly measures the average fault amount and the distributed collective method consistently outperforms the other methods with up to 50% higher detection accuracy.

Statistical methods to estimate vehicle count using traffic cameras

Traffic camera has played an important role in enabling intelligent and real-time traffic monitoring and control. In this paper, we focus on establishing a correlation model for the traffic cameras’ vehicle counts and increase the spatial-resolution of a city’s vehicle counting traffic camera system by means of correlation-based estimation. We have developed two methods for constructing traffic models, one using statistical machine learning based on Gaussian models and the other using analytical derivation from the origin-destination (OD) matrix. The Gaussian-based method outperforms existing correlation coefficient based methods. When training data are not available, our analytical method based on OD matrix can still perform well. When there is only a limited number of cameras, we develop heuristic algorithms to determine the most desirable locations to place the cameras so that the errors of traffic estimations at the locations without traffic cameras are minimized. We show some improvements in the performance of our proposed methods over an existing method in a variety of simulations.

Sensor Network Assisted Collaboration for Pursuit-Evasion Problem

The pursuit-evasion problem represents many important real-world applications. In this paper, we present a new adaptive pursuer collaboration method with the support of wireless sensor networks. We focus on two different evader motions, random movement and active escaping, and develop a moving pattern recognition algorithm based on information obtained by the sensor network. We compare two pursuing strategies, greedy vs. collaboration, and derive their expected capturing times for the two evader moving patterns. By recognizing the evader moving patterns and selecting appropriate pursuing strategies, the new collaboration method is more adaptive, robust, and reduces the capturing time significantly in situations where the number of pursuers is small and the operating field is large

Localize vehicles using wireless traffic sensors

Recently, wireless traffic sensors present themselves as a low cost and non-intrusive alternative to wired traffic sensors. We propose a vehicle localization method that utilizes the signals of the wireless sensors. A vehicle is equipped with a receiver and overhears the geo-tagged packets transmitted by wireless traffic sensors. An onboard computer then computes the distribution of possible vehicle locations using an algorithm based on the principles of particle filtering. In our simulation, the proposed method outperforms the proximity centroid method by an average of 79%.

Cobra: Correlation-Based Content Authentication in Wireless Sensor Networks

The energy-constrained wireless sensor nodes require the designs of energy-efficient content authentication protocols. In this paper, we combine the approaches of cryptography and content analysis to propose a new light-weight content authentication framework Cobra. In Cobra, only a small fraction of transmitted packets are protected with signatures. The unprotected packets are authenticated using the contents of the protected packets and a pre-learned correlation model. Cobra is resilient to four common types of data integrity attacks. Both our theoretical analysis and the experimental results show that Cobra offers high detection rate against aggressive attacks. Compared to three competing methods, it reduces the average data error by up to 60% and the security overhead by an order of one magnitude. As the attacks become more aggressive, Cobras average data error is almost unaffected whereas the average data errors of the competing methods all grow linearly.

A New Method of Using Sensor Network for Solving Pursuit-Evasion Problem

Wireless sensor networks offer the potential to significantly improve the performance of pursuers in pursuit-evasion games. In this paper, we study several sensor network systems, their interaction with the pursuers, and the effect on pursuer performance. We propose a general framework to solve the pursuit-evasion problem and present new centralized as well as distributed methods. Specifically, we address three issues in the design of pursuers based on data provided by the sensor network : a) how to identify evader moving patterns, (b) how to predict the evader locations using different evader moving models, and c) how to choose the most efficient pursuit strategies. We propose efficient algorithms to solve these problems and show that they are effective in reducing the capturing time in our simulations. We also compare the distributed and centralized methods. Experimental results show that the distributed method is efficient and produces solutions close to the centralized method.

Minimizing Location Uncertainty in Deployment of Access Points

In recent years, wireless sensor networks have been applied to several applications of data gathering and target localization across large geographical areas. In this paper, based on a representative sensor network system for lost hiker localization in wilderness areas, we study the problem of sensor/access-point deployment in constrained environments with the goal of minimizing localization uncertainty. We first present theoretical analysis of the problems for different problem attributes. Then we propose an efficient two-phase algorithm to find the access point placement with four different heuristics. In addition, we present experimental results to compare the performance of different heuristics and one heuristic namely divide-merge is shown to outperform all others and achieves solutions very close to the optimal