The research reported here anticipates the future of smart buildings by developing algorithms that categorize the movements of individuals based on such characteristics as motion vectors, velocity vectors, head orientation vectors and predetermined positions. The intended applications include detecting intrusions, helping lost visitors, and changing the artwork on virtual posters to reflect an individual's presumed interests. The vectors we capture represent trajectories in a multi-dimensional space. To make sense out of these, we first segment a trajectory into sub-trajectories, typically based on time. To describe each sub-trajectory, we use primitive patterns of body movement and additional information, e.g., average speed during this interval, head movement and place or object nearby. That is, for each sub-trajectory, we use a tuple of the following form: (interval_ID, body_movement, avg_speed, head_movement, places_passed). Since trajectories may have many outliers introduced by sensor failures or uneven human movement, we have developed a neural network-based pattern extraction subsystem that can handle intervals with noisy data. The choice of these attributes and our current classification of behaviors do not imply that these are the only or best ways to categorize behaviors. However, we do not see that as the focus of the research reported here. Rather, our goal is to show that the use of primitive attributes (low level), neural networks to identify categories of recognizable simple behaviors (middle level) and a regular expression-based means of describing intent (high level) is sufficient to provide a means to convert observable low-level attributes into the recognition of potential intents.
UWB communication is essentially the transmission and receiving of ultra short electromagnetic energy pulses. Short pulses mean wide bandwidths, often greatly exceeding 25% of the nominal center frequency. Modern UWB radio is characterized by very low power transmission (in the range of tens of microwatts) and wide bandwidths (greater than a gigahertz). One of the major applications of Ultra-wide band technology has been for detection and tracking of intruders in different environments. Based on some of our previous work [1,2] we developed a hybrid Ray-tracing/FDTD technique to study the indoor and outdoor propagation of UWB signals. The basic goal of this paper is to describe the experimental and simulation studies that were conducted to locate and track an intruder inside a UWB sensor web system. The sensor was developed using the Time Domain P-200 device and the software was developed using MATLAB. Return scans from UWB devices are analyzed to determine the noise floor and the signal strength. Using the noise floor level a threshold level is set above which the alarm will be triggered to determine the presence of an intruder. The probability of false alarm (PFA) is also determined using the Signal-to-Noise ratio and the threshold. We vary the PFA to lower the false alarm to a minimum level. We also determine the noise statistics of the system using Non-parametric Kolmogorov-Smirnov (KS) test. Using this basic UWB sensor web system we will try to determine the physical dimensions of the intruder and also track multiple intruders on the system.
Proc. SPIE. 4863, Java/Jini Technologies and High-Performance Pervasive Computing
KEYWORDS: Human-machine interfaces, Statistical analysis, Databases, Receivers, Personal digital assistants, Telecommunications, Software development, Distributed computing, Data communications, Global Positioning System
We describe two distinct distributed systems, MeasureMe and GeoPresence. As with all distributed software systems, these require support for communication, coordination, task distribution and persistent shared data. In this paper, we show how the tuple spaces paradigm provides the required communication and coordination, and how its use facilitates our achieving scalability, extensibility, data persistency and ease of software component upgrades.
This document reports the results of the authors' work on understanding the problems associated with dynamic terrain (DT) in networked visual training simulators. Dynamic terrain (construction of emplacements, cratering and repair, etc.) is of substantial military interest as ground-based simulation becomes a common training technology. The basic cost/performance issues of visual simulation are analyzed with regard to the introduction of DT. An overview of current networking (SIMNET) technology for visual simulation is provided, and the difficulty of extending the SIMNET paradigm to dynamic terrain is discussed. An object-oriented representation for terrain is suggested, and its advantages are described. Finally, we consider the implications of dynamic terrain within networked simulation, with particular reference to the problems of scale imposed by the interaction of high speed aircraft flying nap-of-earth (i.e. at treetop level) and low speed ground vehicle simulators.