As part of this effort, the U.S. Army Engineer Topographic Laboratories
has the task of producing a high resolution, high accuracy experimental
digital terrain analysis data base of a 12 square kilometer test site.
The compilation of this data base is being performed on the Computer-
Assisted Photo Interpretatation Research (CAPIR) system. CAPIR is an
ongoing research effort which addresses the issues of digital terrain
data extraction, storage, and exploitation. This integrated system con-
sists of an analytical plotter equipped with stereo-superposition graph-
ics and a geographic information system (GIS) to provide the mechanism
for 3-dimensional data capture, verification, and management.
Objectives for the five year ALV Project involve the development of road
following capabilities, obstacle avoidance and off-road/cross country
traversal. In order to achieve these project goals the vehicle will
require the following capabilities:
l. Perception - the ability to handle and symbolically represent sensor
images from a color video CCD TV camera and ERIM Multispectral laser
scanner.
2. Reasoning - the ability to receive goal directed inputs, control sub-
systems, and derive navigation decisions necessary to achieve this goal.
3. Terrain Knowledge Base Maintenance - the ability to support and
update both a-priori digital terrain information and extracted sensor
data about the surrounding features.
4. Positional Knowledge - the ability to provide and update the 3-
dimensional position of the vehicle over time and location.
This paper describes the construction of the experimental ALV terrain
data base which will serve as the source of a-priori terrain information
for the ALV.
Data Requirements
The ALV project brings a unique perspective to digital mapping. From
this viewpoint, the digital terrain data must support and assist a
dynamic autonomous robot to interpret and navigate the surrounding
environment. Operationally, this increases emphasis on the following
criteria. First, diverse knowledge is required about the encompassing
terrain. Three-dimensional thematic data such as vegetation, cultural
features, roads, soils, surface drainage, landforms, and topography are
important due to their direct influence on mobility and navigation of
the vehicle. Second, the relative horizontal and vertical accuracy of
these data are very important because of the current limitations of
robotics and computer vision/perception. Third, the project requires a
high level of detail about terrain features or structures which might
seriously impede vehicle navigation. This level of detail must not only
be consistent within themes, but also between themes. Thus, individual
478
as
J$5 Ria AU A
bl PM A bat PM d
” oN
