American Institute of Aeronautics and Astronautics
1
THREE-DIMENSIONAL MULTI-OBJECTIVE UAV PATH
PLANNER USING META-PATHS FOR DECISION
MAKING AND VISUALIZATION
Levi Swartzentruber 1
, Jung Leng Foo2, and Eliot H. Winer
3
Virtual Reality Applications Center, Iowa State University, Ames, IA 50010, USA
Military operations are turning to more complex and advanced automation technology for
minimum risk and maximum efficiency. A critical piece to this strategy is unmanned aerial
vehicles (UAVs). UAVs require the intelligence to safely maneuver along a path to an
intended target, avoiding obstacles such as other aircrafts or enemy threats. Often
automated path planning algori thms are employed to s pecify targets for a UAV to fly to. To
date, path-planning algorithms have been limited to providing only a single solution
(alternate) path without further inputs from the UAV controller. This paper presents a
unique platform for decision making in a three-dimensional path planner where multiple
solution paths are generated in the form of meta-paths. The path planner uses Particle
Swarm Optimization (PSO) to generate multiple solution paths based on predefined criteria.
The problem formulation was designed to minimize risk due to enemy threats, to minimize
fuel consumption incurred by deviating from the original path, and takes into account
reconnaissance targets. Using PSO, al ternate paths are generated using B-spline curves,
optimized based on preferences set for the three objectives. The resulting paths can be
optimized with a preference towards maximum safety, minimum fuel consumption, or target
reconnaissance. For each preference, the top five solutions generated by PSO are presented,
for a total of 15 alternate paths. In order to efficiently present these multiple solution paths,
meta-paths are implemented, which is a single summarized representation of all the paths
generated for a particular preference. This allows the decision making process to be
completed in an efficient and organized manner. The problem formulation and solution
implementation is described along with the results from several simulated scenarios.
I. Introduction
ilitary combat of the future will become highly dependent on the use of unmanned aerial vehicles (UAVs). In
recent years, there has been rapid development in UAV technology such as swarm communicat ion, command
and control, and developing usable interfaces 1. The complexity in UAV technology is rapidly growing, and
according to the Department of Defense (DOD) Roadmap 2, by the year 2012 it is estimated that F-16 size UAVs
will be able to perform a complete range of combat and combat support missions. Thus, the ground control station –
the human operator’s portal to the UAV – must evolve as UAVs grow in autonomy. The ground control station must
facilitate the transformation of the human from p ilot, to operator, to supervisor, as the level of interaction with
UAVs moves to ever-higher levels. As humans interface with UAVs at more abstract levels, a UAV will be trusted
to do more 3. To develop and maintain that trust, a human must be able to understand a UAV’s situation quickly.
Future ground control stations will need to provide an operator with situational awareness and quality information at
a glance.
To address the many research issues involved in the command and control that the DOD roadmap requires, a
“Virtual Battlespace” at Iowa State University was created. In this paper, research into the issue of three -
1 Research Assistant, Department of Mechanical Engineering & Human Computer Interaction, Virtual Reality
Applications Center, 2274 Howe Hall, Iowa State University, Ames, IA 50010, USA, Student Member. 2 Research Assistant, Department of Mechanical Engineering & Hu man Computer Interaction, Virtual Reality
Applications Center, 2274 Howe Hall, Iowa State University, Ames, IA 50010, USA, Student Member. 3 Assistant Professor, Department of Mechanical Engineering & Human Computer Interaction, Virtual Reality
Applications Center, 2274 Howe Hall, Iowa State University, Ames, IA 50010, USA, Member.
M
12th AIAA/ISSMO Multidisciplinary Analysis and Optimization Conference 10 - 12 September 2008, Victoria, British Columbia Canada
AIAA 2008-5830
Copyright © 2008 by Levi Swartzentruber. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.
American Institute of Aeronautics and Astronautics
2
dimensional (3D) path planning for UAVs as part of the Virtual Battlespace project is presented. The method
described allows a human operator to focus on selecting an appropriate path from a set of alternate paths produced
by the path planner, easing the decision making process. Using a Particle Swarm Optimization (PSO) algorithm, the
task of generating alternate paths is formulated into an optimization problem consisting of three main components:
1) for UAVs to avoid obstacles such as threats (e.g., surface to air missile sites, tanks, and aircraft), 2) maintaining a
fuel-efficient path to maximize mission range, and 3) minimal deviat ion from the original way-points for
reconnaissance purposes.
In the following sections of this paper, the background of the Virtual Battlespace project is presented. This is
followed by a detailed description of the development and implementation of the 3D path planner using a Part icle
Swarm Optimization algorithm. The process of generating the meta-paths for effective visualization of multiple
alternate paths is also discussed. The resulting paths generated from the planner from multiple simulated scenarios
are then presented, with conclusions and future work discussed at the end.
II. Background
A. Virtual Battles pace
Development of the Virtual Battlespace originated in 2000 when a research team at Iowa State University’s
Virtual Reality Applicat ions Center (VRAC) began work with the Air Force Research Lab’s Human Effect iveness
Directorate and the Iowa National Guard’s 133rd Air Control Squadron. The goal of this preliminary version of the
Virtual Battlespace was to develop an immersive VR system for distributed mission training . Virtual Battlespace
integrates information about tracks, targets, sensors and threats into an interactive virtual reality environment that
fuses the available information about the battlespace into a coherent picture that can be viewed from mu ltiple
perspectives and scales 4, 5
. Visualizing engagements in this way is useful in a wide variety of contexts including
historical mission review, mission planning, prebriefing, post-briefing, and live observation of mission training
scenarios. This system comprises of set of tools developed to enhance the user experience using voice recognition,
immersive visualizat ion, and intuitive interaction using a wireless gamepad controller. In addition to that, a tablet
interface with d irect manipulation where mission specific details can be viewed and updated in real t ime was also
developed. The environment in a large-scale VR environment is shown in Fig. 1.
Figure 1. Virtual Battlespace in the C6 six-wall p rojection system at Iowa State University’s Virtual Reality
Applications Center.
B. Path Planning for UAVs
Real time dynamic path alteration is needed when a UAV is presented with an unexpected threat. For example, a
UAV could encounter an unexpected surface-to-air missile (SAM) site. When this happens the operator must be
alerted to this dangerous situation and be able to quickly re-task the UAV to reduce its threat exposure while
considering other factors such as fuel usage.
It is important to consider the impact of the immersive environment on this process. In a con ventional two-
dimensional (2D) interface, the application would have to find some way to convey a 3D path in the 2D interface or
restrict the path-planning algorithm to a 2D solution; limit ing any alternate paths to changes in direction within the
American Institute of Aeronautics and Astronautics
3
same elevation when in reality an aircraft could also change altitude to avoid threats. This limitation is lifted since
the Virtual Battlespace operates in an immersive virtual reality environment, which allows true 3D interaction. As
such, there is a need for a path planner application that functions in 3D space. With this tool, the operator can focus
on the decision to be made as opposed to inferring the true shape of the path.
There has been extensive research in the area of path planning especially in the artificial intelligence,
optimization, and video game communit ies but most have been restricted to a 2D form 6, 7
. One of the most popular
path planning algorithms in the video game/art ificial intelligence communit ies has been the A* path planning
algorithm. The A* algorithm strength lies in the ability to heuristically judge or value the best path from point to
point. If this cannot be done with reasonable accuracy then the A* method will not be very effective 8. Th is is not
possible considering the dynamic nature of the battlefield and the variable cost of particular parts of the environment
based on differing criteria. Without human intervention, the path planning algorithms must be able to adapt to a
variable mission environment. This has lead to research of great interest not only to UAV control but other fields
such as robotics 9 – 13
.
Without the ability for human intervention of the UAV’s path, the Air Controllers and human p ilots are solely
responsible for maintaining a manageable airspace. Keeping a human in the loop helps prevent catastrophic mistakes
by taking advantage of the human’s ability to handle and process outside information. The human operators can
issue overall object ives and commands to the vehicles under their control. The issuing of obje ctives as opposed to
exact paths can reduce the amount of awareness needed to control an individual UAV. This reduction could result in
more UAVs under the control of a single operator.
Because of the variable cost nature of the types of path planning that will be done with UAVs, a particle swarm
optimization (PSO) method of path planning was developed. To maintain a human input in the decision making
process, several paths are generated by the developed method. The generated alternate paths are represente d by B-
Spline curves to minimize computation, since a simple curve can be eas ily defined by as little as three control points
and this method has been successfully used to model constrained curves 14
.
C. Decision Making in Path Planners
Decision making forms an important aspect of UAV path planning as a human presence is included in the path
selection process to validate proposed actions of the UAV. The path planner operation and output must be designed
to facilitate quick and accurate interpretations of the data so the UAV controller can choose the appropriate
alternative under the tight time constraints of a combat situation. When military officers encounter such situations,
decisions are generally made by creating a vision for the mission outcome, generating a plan which is evaluated and
refined, and finally issuing the specific orders to carry out that plan. Most frequently, it is intuition based on
experience that leads to the best action plan 15
. The path planner generates a number of alternate paths, relieving the
operator of this task, and allows that person to focus on evaluating the paths in light of the mission objective and
choose the best alternative. With more alternatives there is a much greater chance the operator will intuitively
recognize a path that fits very closely with the mission objective.
Intuition and past experience form a core for how the decision will be made. The three stages involved in this
process are model construction, revision, and falsificat ion. Stage one involves making a mode l of the situation. Then
the user seeks out more detailed information to come to a preliminary conclusion which is accepted if it is not
falsified by further information 16
. Such a thinking process would lend itself well to a tiered informat ion display
approach.
Information display techniques can significantly influence the operator’s ability to make that decision well. The
operator will be biased toward interactive v isual information which is absorbed most quickly while textual
informat ion is processed more slowly and given less weight in decision making processes 17
. Proper path display will
leverage this bias by giving key information, such as path optimization criteria and waypoint positions, the most
visual weight.
More information allows for a better situation up to the point where the user can no longer process it quickly.
Several tactics including overview + detail, focus + context and in formation h iding are commonly employed to
prevent information overload as a user investigates the data 18
. Utilizing these types of techniques would allow a
user time to evaluate more alternatives than before in the same amount of time and make a better-informed decision.
Most research on decision making in path planning takes the alternative approach where the goal is to
completely eliminate the need for human input. The methods include dynamic path planning algorithms that
continuously plan short sections of the path 19
and a hierarchical planning structure to develop an overall solution
which is refined as path sections are analyzed in increased detail 20
. Though they are not currently used in UAV path
planning as described in this paper, they could form the basis for UAV control and decision making in the future.
American Institute of Aeronautics and Astronautics
4
The other approach is based on the concept that there can be multip le good alternate paths for a given situation.
One algorithm was designed for dynamic environments using cluster analysis with an evolutionary algorithm. This
would prevent the planner from needing to recalcu late automatically if the dynamics of the situation changed
because there would still be other alternate paths available 21
. The user would select the path that would best fit with
his analysis of the dynamics of the situation.
D. Particle S warm Optimization (PSO)
To facilitate the search for optimal paths, a particle swarm optimizat ion (PSO) technique was used to produce a
large number o f candidate paths for evaluation. PSO is a heuristic optimization method that is based on the
movement of insect swarms introduced in the mid 1990s by Kennedy and Eberhart 22
.
In PSO, an initial randomly generated population swarm (a collection of part icles) propagates towards an
optimal point in the design space, and reaches the global optimum over a series of iterations. Each particle in the
swarm explores the design space based on the informat ion provided by previous best particles. PSO then uses this
informat ion to generate a velocity vector indicating a search direction towards a promising design point, and updates
the locations of the particles.
After reviewing the various current methods and research being done for path planning of UAVs, this paper
presents a new method of path planning in 3D space using the PSO algorithm to generate alternate optimal paths,
using meta-paths for better decision making.
III. Methodology
The path planning process is initialized by determining the starting and ending points of the current path. The
waypoints from the original path are then mapped as reconnaissance targets. The start and end points and any
waypoints in between form the in itial design points of the problem. From this init ial design point, a search space is
defined to scan and locate other UAVS with in range and identify possible threats. The size of the search space is left
open to the user’s judgment, setting it too large will incur a longer computation time, while having a search space
that is too small might cause some UA Vs to be unaccounted for. Fig. 2 shows the process flowchart of the path
planning using PSO.
Once position data of the UAVs within range are obtained, enemy entities are singled out and a 3D threat zone is
generated for each of them. A threat zone is defined as a sphere (a hemisphere for ground vehicles) of radius RT
(user defined) surrounding the obstacle that the path needs to avoid. Threat zones are also generated for non-enemy
(friendly) entities to avoid collision, but with a smaller radius. In addition, reconnaissance zones are also defined as
hemispheres of radius RR. By default, the values for RT are set to be 20,000 feet and RR at 2,000 feet, but can be
changed to suit the UAV controller’s preferences.
Figure 2. Flowchart o f the path planning process using Particle Swarm Optimizat ion.
Identify waypoints of original path and set as reconnaissance targets
Identify enemies or obstacles and generate threat zones.
Run Part icle Swarm Optimizat ion (PSO) to obtain optimal path(s).
Identify start and end points for path planning.
American Institute of Aeronautics and Astronautics
5
A. Cost Formulation
Formulat ion of the optimization cost function begins with the description of a B-spline curve to represent the
path of the UAV. Consider the B-spline curve p0(ui), where ui is a sequence of line segments forming the curve, that
requires re-planning when it violates a threat zone ZT (in red), shown in Fig. 4. A resulting alternate curve path p(ui)
is generated that avoids the threat zone while still attempting to be within the reconnaissance zones (in green) and is
illustrated by the red curve in Fig. 4.
The cost function components also depend on the number of parametric samples (line segments that form the
curve) N that define the resolution of the curve. Here, N is user defined and the value of N brings a trade-off effect
between accuracy of the curve and computational efficiency. The cost function components are summations of the
curve characteristics sampled at the regular parametric intervals,
u0 = 0 and
121 ,1
11 ,....,N-, i
Nuu ii
(1)
The initial solution to begin the optimization process is the original path that breaches the threat zone thus
violating the constraint, illustrated in Fig. 4 by the blue dashed line. A new path can be computed by running the
PSO such that the interior control points (between the end-points) satisfy the constraints. To achieve this, the cost
function needs to accommodate the preferences of safety, reconnaissance missions, as well as fue l efficiency of the
alternate paths. The total cost function is represented by the following components:
C =K1 CT + K2 CL + K3 CR (2)
where, CT is the cost due to proximity of enemy entities and violation of the threat zones, CL reflects the cost
incurred from excessive arc length and deviation from the original path, and CR is the cost incurred by deviating
from the reconnaissance locations.
The constants K1, K2, and K3, in Eq. (2), are component weights that determine the relat ive emphasis of the
various cost components with respect to the overall cost function. Each weight is normalized between zero and one.
If a weight is zero then that particular cost function is unimportant for a particular run. All weights sum up to 1.0 in
total. These weighted cost components are then added together to form the total cost function of a particular path.
Table 1 shows an example of generating a set of three different alternate paths, each with its own preference.
Table 1. Example of component weights used to generate a set of alternate paths.
Threat Weight, K1 Fuel Weight, K2 Recon Weight, K3
Threat Avoidance 0.90 0.05 0.05
Fuel Efficiency 0.05 0.90 0.05
Reconnaissance 0.05 0.05 0.90
Figure 4. Two-d imensional illustration of a simple threat zone avoidance problem.
p0(u)
p(u) ZR
ZT
American Institute of Aeronautics and Astronautics
6
The threat component CT requires a function to determine the distance from a point p(ui) along the curve p(u) to
a UAV inside the threat zone ZT and is denoted here as d(p, ZT). The function will return a positive value if there is a
violation of the threat zone, and negative one otherwise. With this, the threat cost is then defined a s:
d(p(ui), ZT) = (Threat zone radius, RT) – (Distance between point p(ui) and threat) (3)
0)),(p(
0)Z),(p(
0
,)Z),(p( ,
T1
0
T
Ti
iN
i
i
iiTZud
ududCCC (4)
A significant violation of the threat zone will result in a significant increase in the threat component of the cost
function. Since this simple zone violat ion constraint allows many possible solution curves with probably
unacceptably large length, the second component will simultaneously min imize the curve arc length, thus providing
a solution with the best fit possible along the obstacles.
The curve length component of the cost function is computed using a chordal approximation of the total curve
length, L, relat ive to the initial solution obtained from a line connecting the endpoints. The curve length component
is expressed as follows:
path original ofLength :
)(p)(p ,
0
2
0
10
L
uuLLLCN
i
iiL
(5)
The curve length component translates as a difference between the generated path and the original path. This
represents the additional fuel expense from the alternate path. Should the path planner find a shorter route
(regardless if it violates a threat zone), this component will return a negative value, thus turning this component into
a reward rather than a cost. The goal is to generate a new path for the UAV that avoids a threat, with the lowest
additional fuel expense simultaneously.
A third and final component is for reconnaissance, CR, which is a function to determine the distance from a point
p(ui) along the curve p(u) to a particular waypoint location ZR and is denoted here as d(p, ZR). The reconnaissance
component increases the objective function when the path is outside the specified reconnaissance zone. The further
the alternate path is from the waypoints, the higher the cost to the objective function. With this, the reconnaissance
cost is defined as:
d(p(ui), ZR) = (Distance between point p(ui) and center of ZR) - (Reconnaissance zone radius, RR) (6)
0)Z),(p(
0)Z),(p(
0
,)Z),(p( ,
R
1
0
R
i
RiN
i
i
iiRud
ududCCC (7)
B. Meta-path generation
The PSO algorithm is run three times with the weightings described earlier. Normally, only the best path is
taken, disregarding other potentially good paths found in the solution process. To better utilize the information
generated by the PSO algorithm, five of the top paths from each of the three PSO runs are made available to the
UAV operator.
In order to avoid an overload of visual informat ion to the user, the total of 15 alternate paths are presented as
meta-paths. Three meta-paths are created for visualization; one for each group of five paths. As an intermediate step
in the decision-making process, the three meta-paths need to present the operator with enough informat ion to make a
decision on which set of paths to investigate further. The specific data for indiv idual paths is not needed until the
final step. They are designed to preserve the trends of the underlying data in a simplified fashion so the user can
quickly evaluate the alternatives.
Fig. 5 p rovides a process flowchart for how the meta-paths are generated from the PSO data.
American Institute of Aeronautics and Astronautics
7
Figure 5. Process flowchart fo r meta-path creation
Once the PSO path data is acquired, the next step is to select the five paths from the swarm that will be used to
make the meta-path. The first selection criterion is the path’s fitness value. Fitness value and optimality are inversely
related so paths with lower fitness are preferred. By default, the first path selected is the global best, which would
have been returned if only one solution were kept.
To select the next four paths to be returned, a second criterion is needed. As the evolutionary algorithm solves
the problem, solutions will tend to cluster around the optimal point. It is probable that choosing paths solely on
fitness value will lead to duplicate paths being chosen. This would be a case of extra in formation being presented to
the operator with no value being added. The second criterion is added to address this problem.
The second selection criterion is based on the need for the alternate paths to be physically distinguishable. Each
potential path is compared with the paths that have already been selected. Paths are handled in order, moving from
the fittest to the least fit until four are found that meet the requirement. In each comparison the total difference
between the positions of the path points is calculated. If the difference is less than a threshold in any comparison, t he
path is discarded. If the difference is acceptable in all cases, the path is added to the list of alternate paths. The
comparison is laid out in Eq. (8) and Eq. (9):
(8)
(9)
The difference in point locations, diff, is shown in Eq. (8). Th is equation is evaluated n times, where n is the
number of paths already selected. The sum is evaluated for each point in the path, i. Here, pcur is the point in the path
being considered as an alternative path and pn is the nth
path already selected. The x, y and z represent the x, y, and z
coordinates for the ith
point for the respective paths.
The difference values, diffn, for the path under consideration are then compared with a user defined constant. The
constant value, C, is the number of points in the paths mult iplied by a scale factor. The factor is experimentally
chosen to provide good visual separation between paths.
For each of these selected paths, a relative fitness value is computed. With the path planner returning mult iple
paths, sub-optimal paths are being presented to the operator. The user needs a measure of how far the solutions are
from optimal to make an intelligent selection. Relative fitness is a ratio of the fitness of a given path to the fitness of
the best path found through the evolutionary algorithm. This is chosen over presenting the raw fitness values
because the ratio presents the operator with a standard baseline for what is good when fitness values can vary
significantly between test cases.
The calculation for relative fitness is shown in Eq. (10). The calcu lation is performed for each of the five
alternate paths, i. The fitnesspath i is the fitness value for the ith
path and the fitnesspath 1 is the fitness value for the
optimum path which is saved as path one.
(10)
American Institute of Aeronautics and Astronautics
8
In relative fitness, the optimal path has a value of 100. All other paths have larger fitness values than the optimal
solution and will therefore have relat ive fitness values greater than 100. The larger the relative fitness value, the
more cost the path incurs and the farther it is from optimal.
The third step is to take these five paths and create a single visual representation for them, a meta -path, which
would be displayed to the UAV operator. The meta-paths need to convey generally what the underlying paths look
like so the user can make an informed decision and is not surprised by the individual path characteristics when they
are displayed. As seen in Fig. 6, a meta-path is a bent, variable diameter tube. It is visualized half open so the user
will be able to see all three paths even when one is in front of or overlaps another. The geometry of the meta -path is
defined by the bend locations and a radius value at those locations.
Below the meta-paths in Fig. 6 are the 15 individual paths in their groups of 5. It is clear that presenting all 15 of
these paths to the user at once would cause informat ion overload. It can also be seen how the individual paths, when
mentally grouped, would intuitively be represented in a manner similar to the meta-paths used here.
Figure 6. Meta-path representation with same v iew of all 15 alternate paths
This similarity is created by constructing the geometry of the meta-path from the data for the five paths it
represents. The bend locations, or waypoints, of the meta-path are the averaged locations of the waypoints for the
five individual paths. Eq. (11) demonstrates how this is accomplished. For the meta-path, there are j points with x, y,
and z components. The meta-path point is pmeta-path. The individual path points are pi where i is ranges from 1 to 5 for
the 5 paths.
(11)
American Institute of Aeronautics and Astronautics
9
The radius of the meta-path at a bend location is defined as the average separation, in three dimensions, between
the meta-path point and the five individual waypoints. Using th is radius value, the meta-path occupies roughly the
same volume visually as the five individual paths. This calculation is shown in Eq. (12):
(12)
The radius is a value calcu lated at each waypoint, j, of the paths. Again, it is necessary to average the calculation
over all 5 individual paths, i. The square root term is a distance calculation in three dimensions. As before, pmeta-path
is the meta-path point and pi is the path point. The xj, yj, and zj are the x, y, and z components of the position of the
jth
point. A larger radius value means there is more physical distance between the path points at a given waypoint
and results in a larger meta-path. The reverse is also true. Since the beginning and ending points are the same for
all paths, each meta-path comes to a point at both ends.
C. Path Selection
The second aspect of path planning is the decision making process performed by the operator. Meta -paths are
designed to give this individual a wider base of options to consider when tasking the UAV. In doing this, another
layer was added to the decision making process, which is shown in Fig. 7. The operator must select a meta -path, an
individual path, and then confirm that selection for the alternate path to replace the current path. It is possible to
navigate freely between the indiv idual path mode and meta-path mode so the operator can investigate all 15
possibilit ies if that is necessary to find a suitable solution. The operator also reserves the ability to quit at any time
and use the original path tasked to the UAV.
Figure 7. Flowchart for path selection process
American Institute of Aeronautics and Astronautics
10
IV. Implementation of PSO Path Planner into Virtual Battlespace
The purpose of an immersive command and control station is to permit the operator to focus on the overall
mission status. As the number of aircraft under an operator’s control increases, it becomes impossible to constantly
monitor and manage every aircraft. To facilitate this, an alert subsystem was developed as part of the Virtual
Battlespace to alert the operator to any issues. The alert subsystem plays a vital ro le in reassuring the operator that
when UAVs run into situations that require user input, the operator will be made aware of them.
The path planning process begins when a threat is detected by the alert subsystem of Virtual Battlespace, which
prompts the controller for a decision on the next action. The controller can either ignore the alert or choose to
inspect it. The process ends here and does not execute the path planner if the controller decides to ignore the alert.
The alert subsystem, seen in the right image of Fig. 8, notifies the operator of the presence of an alert and when
the operator chooses to examine an alert posted by a threatened UAV, the operator will see a variety of
automatically generated path options. These path options will appear at a distance corresponding to a default value
of 30 seconds ahead of the UAV’s current position and reengage with the path when in a safe region. This lead -time
can be adjusted by an operator. These points on the old path are used as the start and end points of the path planner.
All relevant threats, reconnaissance targets, and the start and end points are passed to the path planner to calculate
new candidate paths.
Figure 8. Illustration of a path in the Virtual Battlespace environment (left ) and a threat alert d isplay (right).
Several distinct scenarios were used to test the developed path planner. For each scenario, three alternate sets of
paths were generated. For the purpose of this paper the parameter settings used were those in Table 1. However, an
operator can adjust these weights if additional paths for rev iew are desired. The first alternate path has preference
towards reconnaissance locations. This is significant when the importance of a UAV’s current mission or future
mission may demand that the UAV stay as close as possible to its original waypoints . The second path is weighted
towards threat avoidance, and the last path makes a preference for fuel efficiency (minimal fuel expense for an
alternate path). The operator also has the option to vary these parameters to fit the mission objective.
The generated paths are represented in different colors with unique labels for easy identification and inspection,
and are represented as follows:
Table 2. Color and label representations of generated alternate paths.
Color Label
Reconnaissance Blue A
Threat Avoidance Green X
Fuel Efficiency White Y
A. Simulated Test Scenario #1
The first scenario test is one where three threats are situated in close proximity to each other and are in the way
of the UAV’s original path, as can be seen in the right view of Fig. 9. The threats are represented as red spheres and
their corresponding threat zone is represented as a set of red rings. After running the PSO path plann er, the resulting
meta-paths were generated and displayed as shown in Fig. 9.
American Institute of Aeronautics and Astronautics
11
Figure 9. Right and top meta-path views for the first scenario.
The meta-paths in Fig. 9 reveal some general trends of the individual paths used to create them. For instance, the
larger spread of the threat avoidance meta-path suggests its paths are more varied than the reconnaissance or fuel
efficiency paths. Its paths also tend to stay farther away from the threats while the reconnaissance paths stay closer
to the original path and the fuel efficiency path tries to take the shortest route. The operator has the option to
investigate any of the meta-paths or use the return functionality to look at all three groups of paths.
a) Reconnaissance alternate paths
c) Fuel Efficiency alternate paths b) Threat avoidance alternate paths
Figure 10. Alternate paths displayed when respective meta-path is selected
American Institute of Aeronautics and Astronautics
12
Each of the meta-paths breaks down into its five individual paths when selected. Fig. 10 shows the alternate
paths for each of the choices. In each case, the best path is initially highlighted. The highlighted path is the currently
selected path and its relative fitness value is displayed on the screen. Relat ive fitness provides a numerical means for
the operator to evaluate the alternate paths in addition to the visual inspection. The optimum value is 100 and less -
than-optimal solutions have larger values. Thus, the operator can weigh the optimality of the solutions against their
physical layout to make an informed decision about how to reroute the UAV.
Table 3 shows all the relative fitness values for each of the paths in this test case. As mentioned, the first path is
the basis for the comparison and will always have a value of 100. It is apparent that the reconnaissance paths have a
much s maller range of values than the threat avoidance paths. This is due in large part to the problem formulation
within the PSO algorithm. Reconnaissance paths are constrained to move in a relatively s mall area since they need
to get close to the waypoints. Threat avoidance paths simply want to stay away from the threats and so have a much
wider range of possibilities and therefore a wider range of values. Fuel efficiency paths are most affected by path
length and since all paths begin and end at the same spot and have the same number of waypoints, large differences
are not expected in this group.
Table 3. Relative fitness values for all 15 paths returned by the planning algorithm
Path Relative Fitness Path Relative Fitness Path Relative Fitness
1 100.0 1 100.0 1 100.0
2 100.7 2 106.5 2 103.3
3 100.9 3 118.2 3 103.5
4 101.1 4 125.9 4 104.0
5 101.3 5 142.7 5 104.8
Reconnaissance Threat Avoidance Fuel Efficiency
The threat avoidance case was selected to investigate in more detail in this scenario. The optimal path, on the
left in Fig. 11, travels around the upper left side o f the threat dome. A second interesting goes essentially straight
over the top of the threat dome as shown in the middle of Fig. 11. Its cost is 125% of that for the optimal path. This
option would be attractive for keeping the UAV over the initially created waypoints. A third interesting path curves
through the upper middle of the thread dome. This path avoids the very steep climb and descent of the middle path
in Fig. 11 but has a higher relative fitness since it passes closer to the threats. These are characteristic of the
decisions and tradeoffs presented to the UAV operator. With the additional paths, it is possible to find a path that
better fits what the operator would like the UAV to do. Fig. 12 presents the finalized path, which was the best path
from this group.
Figure 11. Three of the alternate paths for threat avoidance
American Institute of Aeronautics and Astronautics
13
Figure 12. UAV path updated with chosen alternate path
B. Simulated Test Scenario #2 and #3
Figure 13. Test Scenario #2
a) Threats and meta-paths for the scenario b) Optimum fuel efficiency path highlighted
c) A lternate fuel efficiency path selected d) UAV path updated with selected path
American Institute of Aeronautics and Astronautics
14
The second scenario path features two threats positioned together on one side of a bend in the UAV’s path. The
threats can be seen in Fig. 13 along with the meta-paths. The meta-path for reconnaissance cuts closer to the threats
to stay close to the original path. The threat avoidance and fuel efficiency take almost identical paths, bulging away
from the threats, though perhaps not moving out as far as would be expected as these paths are still well within the
threat dome.
For this scenario, the fuel efficiency paths were investigated. As seen in Fig. 13, these paths take a similar
trajectory up and to the right of the original path to avoid the threats. The main difference between the paths is their
separation from the threat. The optimal path in this case is the one that is positioned furthest to the right. In this case,
a solution from the middle of the group was chosen. It had a slightly higher fitness value but deviated less from the
original path. This path, the optimal path, and the final UAV course are shown in Fig. 13.
The third scenario includes two threats closely spaced to one side of an essentially straight path as shown in Fig.
14. The meta-paths exhibit expected behavior in this situation. The threat avoidance paths lead around the left side
of the threat dome. Fuel efficiency and reconnaissance paths go straight to take the shortest path with a moderate
increase in elevation.
The reconnaissance paths are investigated in this scenario. As seen in Fig. 14, these paths essentially stack on
top of each other. Selecting between these paths becomes a matter of how much space should be maintained
between the threats and the UAV. In this case, the middle of the five paths was chosen and is pictured in Fig. 14.
Figure 14. Test scenario #3
a) Front view of meta-paths b) Right view of meta-paths and threat dome
c) Reconnaissance alternate paths d) UAV path updated with selected path
American Institute of Aeronautics and Astronautics
15
V. Conclusion
A three-dimensional path planner was developed to intelligently generate a set of alternate paths to be s elected
by an operator of a UAV. Based on the top generated paths for each preference, a meta -path is generated. Meta-
paths allow for effect ive visualization of mult iple paths for decision making purposes. From the test cases using
different simulated scenarios, the three-dimensional PSO path planner successfully generated alternate paths
satisfying its respective objective as set in the component weight parameters. The three objectives to either maintain
as much of the original path as possible for reconnaissance purposes, to ensure maximum safety, or to maintain
maximum fuel efficiency were successfully satisfied. Most importantly, these paths were generated in real time to
allow for efficient decision making by the UAV controller.
By performing the path planning in three-dimensional space, the solution paths presented are more realistic to
what UAVs are actually capable of performing. The option of selecting a particular path from a set of solutions
ensures that the human factor is still part of the decision making process. With multiple views to evaluate the
generated alternate paths allows the operator to make informed decisions based on the current mission objective.
The feedback that has been received from experts within the field of UAV control indicat es that this is a relevant
and interesting concept that warrants further investigation.
Pending implementations to the existing path planner involve the addition of the functionality to be added is to
develop a path planner that is dynamic in nature to incorporate time as a variable when alternate paths are being
generated, since the position of threats could change in the future of the alternate path. Another implementation in
progress is the inclusion of terrain informat ion in the path planning process.
Acknowledgments
This research was supported by the Air Force Office of Special Research Labs.
References 1Barry, C.L., and Zimet, E. – UCAVs Technological, Policy, and Operational Challenges, Defense Horizons, Center for
Technology and National Security Policy, National Defense University, No. 3, October 2001. 2US Department of Defense, Unmanned Aerial Vehicles Roadmap 2002-2027, December 2002, Available:
http://www.acq.osd.mil/usd/uav_roadmap.pdf. 3Miller, C. – Trust in Adaptive Automation: The Role of Etiquette In Tuning Trust via Analogic and Affective Method. 1st
International Conf. on Augmented Cognition, Las Vegas, NV; July 22-27, 2004. 4Walter, B.E., Knutzon, J.S., Sannier, A.V., and Oliver, J.H. – Virtual UAV Ground Control Station. AIAA 3rd “Unmanned
Unlimited" Technical Conference, Workshop and Exhibit, Chicago, IL, September 2004. 5Ruff, H.A., Calhoun, G.L., Draper, M.H., Fontejon, J.V., and Guilfoos, B.J. – Exploring automation issues in supervisory
control of multiple UAVs. Human performance, situation awareness, and automation: Current research and trends, Mahwah,
NJ, Lawrence Erlbaum Associates, Inc, Vol. II, pp. 218-222, 2004. 6Davis, L. – Warp Speed: Path Planning for Star Trek: Armada. AAAI Spring Symposium AAAI Press, Menlo Park, CA 2000. 7Frazzoli, E., Dahleh, M., and Feron, E. – Real-time Motion Planning for Agile Autonomous Vehicles. Journal of Guidance,
Control and Dynamics, Vol. 25 no. 1, 2002. 8Stout, B. – Smart Moves: Intelligent Path Planning. Game Developer Magazine, July 1997. 9Li, S., Boskovic, J.D., Seereeram, S., Prasanth, R., Amin, J., Mehra, R.K., Beard, R.W., and McLain, T. – Autonomous
hierarchical control of multiple unmanned combat air vehicles (UCAVs). American Control Conference, 2002. 10Beard, R.W., McLain, T.W. – Multiple UAV cooperative search under collision avoidance and limited range
communication. 42nd IEEE Conference on Decision and Control, 2003. 11Schouwenaars, T., De Moor, B., Feron, E., and How, J. – Mixed Integer Programming for Multi-Vehicle Path Planning.
European Control Conference, 2001. 12Mitchell, I., Shankar Sastry, S. – Continuous Path Planning with Multiple Constraints. Technical Report 2003, Available:
www.eecs.berkeley.edu/Pubs/TechRpts/2003/4157.html. 13Yang, H., Zhao, Y. – Trajectory Planning for Autonomous Aerospace Vehicles Amid Known Obstacles and Conflicts.
Journal of Guidance, Control, and Dynamics, Vol. 27, No. 6, 2004. 14Malhotra, A., Oliver, J.H., and Tu, W. – Synthesis of Spatially and Intrinsically Constrained Curves Using Simulated
Annealing. Journal of Mechanical Design, March 1996. 15Thunholm, P. – Under Time Pressure: An Attempt Toward a Prescriptive Model of Military Tactical Decision Making.
How Professionals Make Decisions, Florence, KY, Routledge, pp 43-56, 2005. 16Shrinivasan, Yedendra, and van Wijk, Jarke – Supporting the Analytical Reasoning Process in Information Visualization.
Conference on Human Factors in Computing Systems, New York, NY, 2008, pp. 1237-1246. 17Lurie, N.H., Mason, C.H. – Visual Representation: Implications for Decision Making. Journal of Marketing, Vol. 71, pp.
160-177, January 2007.
American Institute of Aeronautics and Astronautics
16
18Tominski, C., Schulze-Wollgast, P., Schumann, H. – 3D Information Visualization for Time Dependent Data on Maps. 9th
International Conference on Information Visualization, 2005, pp. 175-181. 19Estlin, T., Gaines, D., Chouinard, C., Fisher, F., Castano, R., Judd, M., Anderson, R.C., Nesnas, I. - Enabling autonomous
rover science through dynamic planning and scheduling. Aerospace Conference, IEEE, pp. 385-396, 5-12 March 2005. 20Chandler, P.R., Pachter, M. – Research issues in autonomous control of tactical UAVs. American Control Conference,
1998. 21Hocaoglu, C., Sanderson, A.C., - Planning Multiple Paths with Evolutionary Speciation. IEEE Transaction on Evolutionary
Computation, June 2001, pp. 169-191. 22Eberhart, R.C., and Kennedy, J. – A New Optimizer Using Particle Swarm Theory. 6th International Symposium on Micro
Machine and Human Science, Inst. of Electrical and Electronics Engineers, Piscataway, NJ, 1995, pp. 39-43.