This project is based upon a NASA Small Business Innovative Research (SBIR) solicitation. SBIR solicitations encourage innovative research and are usually formulated in a somewhat open-ended manner. A subtopic of 08.03 Technology for Search and Rescue sought innovations in the area of Search and Rescue (SAR) data resource management to increase the efficiency and accuracy of locating an individual trapped in some sort of hazardous situation. The team defined its project proposal by considering several systems issues. Our initial project proposal focused on the answers to four key questions:
The system will increase the efficiency and accuracy of locating features in the environment. For Search and Rescue (SAR), the feature to locate is an individual trapped in some sort of hazardous situation. The system will provide technological aid in the search and rescue mission. The goal is to take the "search" out of "search and rescue". We will focus on beaconless Search and Rescue, where the individual in distress or other target of interest does not have the capability to emit an emergency distress signal. Thus, it is a requirement that we use remotely sensed data. There are many types of remote sensors and platforms, from numerous orbiting satellite platforms or aircraft. This data is generated by different organizations, such as:
The instrumentation on available sensing platforms of interest, i.e. space-based or air-based, were not developed with the purpose of integrating their data sets with other platforms. There currently exists a plethora of different imaging libraries taken with heterogeneous sensing equipment. Some of the differences between image objects are inherent in the object itself, for example, the number of pixels that the instrumentation has or the type of sensing the instrument was designed to perform, such as infrared sensing or Red-Green-Blue color sensing. Other differences between image objects are an artifact of the different techniques used to store, query, access, retrieve, and display the images. Data management and mapping techniques will be a large portion of an enhanced SAR capability.
This heterogeneous data will be used to enhance existing SAR capabilities. Many technologies are needed to effectively use these data sources in actual SAR operations and generate information that is useful in planning speedy and effective SAR operations. Some of the interesting technologies that might be applied in a global SAR Mission architecture are:
This project will focus on SAR database access where there is a requirement for new technologies in information resource management. Civilian SAR organization do not fly their own satellite and many do not have funding for airborne reconnaissance aircraft. Therefore, the SAR organization will need to access the databases of many different, heterogeneous organizations. Tools will be needed to aid in the efficient and effective search, retrieval, and interpretation of the retrieved sensor data. This project will focus on three technologies to enhance SAR database access:
The system, as currently envisioned, consists of a layered architecture,
as shown in Figure 1. The bottom layer contains raw sensor data,
for example data sensed from an orbiting spacecraft or low flying
airplane. The middle layer contains the Geographic Information
System (GIS). The GIS provides a standardized data format for
sensor data from satellites and analysis capabilities to incorporate
other data sets into the baseline satellite data set. Sample data
sets might include slope, or vegetation characteristics of an
area. GIS permits a user to analyze the relationships between
different data sources. The expert system (ES) layer resides on
top of the GIS layer. The ES layer allows an inexperienced or
novice user to perform sophisticated analysis and correlation
of accessed data sets. The ES layer may be able to suggest new
data sets to access for the user, as well as provide suggestions
for search priorities. In addition to managing the information,
these technologies can be used as an aid to prioritize the search
process. In other words, the search area may be too large to search
its entirety, and methods must be used to prioritize the search
procedure.
Although the project will not focus on a key enabling technology, the interagency real time or near real time database access is shown as a dotted line in Figure 1. The proposed architecture will support this function. The interagency database access will occur at the bottom layer of the architecture, where raw satellite data images are accessed and the associated image description data is stored. The image description data contains the information needed by the GIS to accurately interpret the satellite image.
There are difficult practical issues of systems integration in this project, because the data collections are large and the systems being integrated are complex. GIS packages, knowledge base systems, and statistical data analysis are three examples of activities which are usually relagated to separate, large computer systems.
Each generally has more functionality than is required by a single user, but different users generally have different system and information requirements to solve their problem. A major problem to address in this system design project is the development of a common user interface across the different software tools that are integrated into our final solution. Therefore, user interface standards will be a major part of our research effort.
The second system challenge imposed by this project is the accessibility
and operability of large, heterogeneous, and distributed databases
of raw image data. The heterogeneous nature of these image databases
is a function of both the inherent quality of the image object
itself, and the methods used to store, query, access, and retrieve
the image. Some characteristics of the images are a function of
the instrumentation which was used to generate the image. Examples
of instrumentation features might be:
The methods or database systems used to implement image libraries can also contribute to the heterogeneous nature of these distributed databases. Relational database techniques and software were designed and optimized for alphanumeric data which is laid out in related tables. This problem may be more suited to a combination of object-oriented and relational database techniques to classify and store image "meta-data". The second major endeavor of our work will be research into applicable database standards for these distributed image libraries.
There is currently no integrated solution available to assist SAR organizations in their rescue operations although a few specialized tools do exist. There is no comprehensive capability to access image data, process that data in a GIS package, and provide assistance in planning and implementing a rescue operation. Since a large number of people end up getting lost and needing to be rescued, there would be an increase in the efficiency with which SAR organizations are able to carry out their missions. (By one estimate, there is an average of one search and rescue mission in New Mexico every 36 hours. See NM Search and Rescue). This, in turn, would save taxpayer dollars.
It is anticipated that this system would form the basis of a nationally
available SAR system for authorized users. The system is designed
with civilian agencies such as the Federal Emergency Management
Agency (FEMA), state and local governments, and privately sponsored
SAR organizations.