INTRODUCTION

Next decade may very well be the "the Decade of the Sensor". The proliferation of inexpensive sensors for climate, environment and position and the wide availability of remote sensing data from satellites is starting to dramatically change the way data is gathered and displayed. Because of the high computational nature of many GIS tasks, special thought is needed to integrate real-time data into GIS based visualization and analysis systems.This paper focuses on the computational constraints and the dataflow of real-time data in such a system. A prototype system has been constructed utilizing Esri Map Objects in C++ .

RESEARCH LONG RANGE OBJECTIVES

To study and develop software methods and objects for incorporating real time remote sensing data from a variety of sources into Geographic Information Systems (GIS) for improving command and control response time and effectiveness. To develop methodologies and software structures to facilitate the adding of real time sensor data into geo-referenced software systems.

RESEARCH SPECIFIC GOALS

To (1) examine and categorize forms of available real time sensor data into various types and data uses, (2) derive methods and models for incorporating appropriate data into GIS systems, (3) do systems engineering, analysis modeling, and the object oriented design of the proposed database oriented GIS system and (4) develop a prototype Real Time Sensor Based GIS (RTSB-GIS) system for proof of concept and methodology verification.

TECHNICAL BACKGROUND

Currently, our ability to collect intelligence and sensor data has outstripped the ability to disseminate and utilize it. One possible solution for part of this problem is to incorporate the incoming real time sensor data into a data stream that can be handled by existing GIS systems. An object oriented approach to this problem is proposed to allow high levels of software re-use when changing sensor types or data evaluation and interpretation modules. This data can then be distributed via real-time links to provide tactical commanders with a common tactical picture. These real-time links may also allow mission planning while the mission is in progress.

SOFTWARE ARCHITECTURE

The software being developed is called the Spatial Analysis System (SAS). SAS is the integration of a specialized GIS system with the following main features:

Highly reconfigurable real time display
Spatial analysis features
Self reconfiguration driven by user input and input from sensors

A screen shot of the main program is given below:

The overall program architecture is shown below:

The Graphics User Interface (GUI) main features are:

User Display
Map, Sensors, Layers, etc.
User Control
Commands Display Configuration
Commands Sensor Filters
Commands Security

Based on Map Objects
GIS Function Requests are requested by Incoming Sensor Data
GUI Input
Primary Functions
Spatial Data Retrival
Overlays

Sensor Data
Command Storage
Database Management Routines
Based on MDB Relational Model
Tables, Relations Programmable by Sensors and GUI

Security is vitally important to a "Connected" system
Uses a MOM (Meta Object Model) Provably Secure Communications Layer
Sensor Data Functions
Filter, Discard, Interpolation, etc.
Sensor Meta-Data
Formating
Units
Map Locations
Display Reconfiguration

From a system perspective, SAS has significant issues with processor loading. As the system is designed to be moved to a severe resource limited environment (i.e. a palmtop computer), substantial attention needs to be paid to processor loading and thread contention. SAS is built with three threads running simultaneously as shown below:

These threads are dynamically balanced by the program depending on loading based on processor workloads, user commands and by sensor commands.

FUTURE DIRECTIONS

The SAS program is being moved to Windows CE 3.0+ in an attempt to move the system to a smaller form factor computer. The GIS portion of this program will need to be rewritten in order to use less resources.

CONCLUSION

SAS type applications are becoming possible to implement on desktop and on palm top computers. One of the major difficulties with this process is load balancing and resource allocation. Displays and storage space are becoming a significantly smaller problem. As internet (or network) aware applications become more common, there will be a strong drive to move GIS applications to smaller form factor computers. The ability of a properly designed object oriented design to be self modifying allows many new applications.

ABOUT THE PRIMARY AUTHOR

Dr. John Shovic has been involved in technology and software for twenty five years. He has been active in software and hardware design for most of that time. Dr. Shovic is a founder of two companies, Advanced Hardware Architectures, a semiconductor company and TriGeo, Inc., an Idaho based GIS company. He currently is on the Computer Engineering Faculty at Washington State University and directs a new laboratory for Computational GIS located in the Electrical Engineering/Computer Science department.