A study on Artificial Potential Fields

Kadir Firat Uyanik

KOVAN Research Lab.
Dept. of Computer Eng.
Middle East Technical Univ.
Ankara, Turkey
[email protected]

Abstract Having dealt with a very basic local sensing based

path finding algorithm -tangent bug - in the first homework, we
will focus on again a very simple reactive path planning method,
artificial potential fields(APF). APF is a reactive approach since
the trajectories are not planned explicitly but obtained while
executing actions by differentiating a function what is called
potential function. A potential function is differentiable realvalued function U : Rm R. The output of a potential
function can be taken as the energy and its negative gradient
as the net-force acting on the robot. In this report, two different
potential functions are presented; attractive-repulsive potential
function, and navigation function. These two functions are
tested in a simulation environment with various configurations.
Experiments are done on the robocup small-size league robot
soccer platform that I have been started developing with the
first homework. At the end, we see that both methods have
some limitations and innate problems.

Fig. 1. From left to right: maximum, saddle, and minimum critical points
are shown where figures at the top are the functions and the below are the
gradient vectors of them.

I. INTRODUCTION
The main idea behind artificial potential field methods is
finding a function that represents the energy of the system,
and generating a force on the robot so that the energy of
the system is minimized and reach its minimum value,
preferably only, at the goal position.
Therefore, moving to the goal position can be thought as
moving robot from a high-value state to a low-value state by
following a downhill path. Mathematically speaking, this
is the the process of following negative gradient, or gradientdescent, i.e. c. (t) = U (c(t)). The motion terminates as
soon as the gradient vanishes, that is U (q ) = 0 where
q is called the critical value of U . This critical point is
either maximum, minimum or saddle point as it shown in
figure 1.Although the maximum is not that critical as soon
as robot doesnt start movement from this point, but the
local minimum points are the real concern while designing
a potential function. Here, one assumption is that the saddle
points -being unstable points- do not cause any problem due
to the robots inertia and other dynamics.
In this document, two main approaches, additive attractive/repulsive potential, and navigation potential functions
are to be investigated. The rest of the article organized as
follows: first, a brief mathematical background of potential
functions are given, then the experimental framework is
represented. Lastly, experimental results, pros and cons of
the two methods are given in the discussion section, and it
ends with conclusion.

II. POTENTIAL FUNCTIONS

Two different potential functions are examined in this
report, additive attractive/repulsive potential functions and
navigation potential functions.
A. Additive Attractive/Repulsive Potential Functions
The intuition behind the additive potential functions is
attracting the robot to the goal position while repelling the
robot from the obstacles by superposing these two effects
into one resultant force applied on the robot. This potential
function can be constructed as U (q) = Uatt (q) + Urep (q),
that is the sum of the attractive and repulsive potentials.
1) The Attractive Potential: The essential requirement
in construction of the attractive potential is that Uatt
should be monotonically increasing with the current distance
d(q, qgoal ) of the robot to the goal, qgoal . Two common Uatt
functions are the ones having conical and/or quadratic forms.
In the experiments, combinations of the two are used since
the conical functions suffer from the discontinuity in the goal
position, that is division-by-zero problem in computational
sense.
{
1
d2 ,
d 6 d?goal
Uatt (q) = 2 ?
(1)
2
dgoal d 12 (d?goal ) , d > d?goal
where d = d(q, qqoal ), is the attraction gain, and d?goal is
the threshold after which quadratic function changes to the
conical form.

The gradient of the function (1) is obtained as;

{
(q qqoal ),
d(q, qgoal ) 6 d?goal
Uatt (q) =
(qqqoal )
?
dgoal d(q,qgoal ) , d(q, qgoal ) > d?goal

(2)

2) The Repulsive Potential: The behavior of a proper

repulsive potential function should look like the behavior
of a tightened spring, or the magnets getting closer to each
other so that obstacle avoidance can be satisfied. In this study
following formula is utilized,
{
2
1
( 1 1? ) , D(q) 6 Q?
(3)
Urep (q) = 2 D(q) Q
0,
D(q) > Q?
Derivative of the repulsive potential function(3) is the
repulsive force which is,
{
1
( Q1? D(q)
) D21(q) D(q), D(q) 6 Q?
Urep (q) =
0,
D(q) > Q?
(4)
where Q? is the distance threshold for an obstacle to create
a repulsion effect on the robot, and is the repulsion gain.
The gain parameters (i.e. and ) and the threshold
parameters(i.e. Q? and d? ) are set empirically.
The total repulsive potential field can be obtained by
summing up the potentials caused by all of the obstacles,
Urep =

Fig. 3. A screenshot from the Webots simulator, while robot is tested in

a three-obstacle between robot and the goal case.

As in the case of attractive/repulsive potential functions,

navigation functions are used to move the robot from an
arbitrary configuration to the goal configuration in the negative gradient direction of the potential at that instant. The
navigation potential functions can be constructed as follows:
=

(d(q, qgoal ))2

1

((q) + d(q, qgoal )2 )

(5)

where (q) is a repulsive-like function calculated as

Urepi (q)

i (q) = (d(q, qi ))2 ri 2

k=1

and the movement is realized by following the negative

gradient of the sum of attractive/repulsive potentials or
simply the following vectorial sum;
U (q) = Uatt (q) Urep (q)
In this study, the translation velocity assignment is simply
done by utilizing U (q).
B. Navigation Potential Functions
Due to the local minima problem in the additive attractive/repulsive potential fields, many other local-minima free
potential field methods are proposed such as wave-front
planner [6][7]. However these methods have discretization
problems and thats why they become computationally intractable for high dimensional and large configuration spaces.
To solve the local minima problem, a special function to be
constructed which has the only minimum at the goal position.
These functions are called navigation functions, formally
defined in [8][9].
A function :Qf ree [0, 1] is called a navigation
function if it
k
is smooth (or at least C for k > 2)
has a unique minimum at qgoal in the connected component of the free space that contains qgoal .
is uniformly maximal on the boundary of the free space,
and
1
is Morse
1 Detailed

discussion can be found in [1].

for each obstacle having a radius of ri . The critical point

in choosing Bi (q) is that QOi = q | i (q) 6 0; this means
i (q) is negative inside an obstacle and it is positive outside
in the free space, which requires
0 (q) = (d(q, q0 ))2 r0 2
where q0 is the center and r0 is the radius of the sphere
world.
In this report, it is assumed that the configuration space is
bounded by a sphere centered at q0 and the other obstacles
are also spherical centered at qi .
The in the equation (5) is necessary for bounding the
navigation function so that it has 0 at the goal and 1 on the
boundary of any obstacle.
The on the other hand, has an effect of making the
navigation function has the form of a bowl near to the goal.
Increasing causes the other critical points(local minimum
etc.) shift toward the obstacles and resulting in a negligible
repulsive behavior compared to the overwhelming influence
of the attractive field. The effect of increasing is shown in
the figure2, for the case of having three obstacle in front of
the robot and a goal position beyond the obstacles as it is
illustrated in the figure 3.
Once a proper value is decided, robots configuration
or -in this study- velocity is updated with the negative
gradient of the navigation potential(equation (6)) function
similar to the whats been done in the attractive/repulsive
potential functions.

(
=

2dd[((q) + d2 ) )] 1 (q + d2 ) 1 ((q) + 2d21 d)

2

(6)
2
((q) + d2 )
)
N (
N
where d = d(q, qgoal ) and (q) = i=0 i (q) j=0,j6=i i (q) . i (q) = 2(q qi ) for i > 0, 0 (q) = 2(q q0 ).

Fig. 2. From left to right and top to bottom, is changed between 1 and 9, and navigation potential function is calculated for the sphere-world by
sampling the world for the specific configurations(viz. 2D positions) varying in polar coordinates r = [0.01, 2.0] and = [0.01, 2] with the discrete
sampling steps of 0.01. As it is clearly shown that doesnt effect the navigation potential considerably after a certain value, which is = 5 in this case.

III. E XPERIMENTAL F RAMEWORK

In this study, I have used Webots[5] robotics simulator
by adding ROS (Robot Operating System,[4]) support to the
framework that I have been developing for RoboCup Small
Size League robot soccer competition. The robot design is
done in VRML (Virtual reality modeling language) which
has four asymmetrically distributed omniwheels enabling
robot to translate and rotate at the same time while following
a linear path.
Below is the kinematic equation for the robot that is used
navigate robot as if it is a point robot but having a specific
radius:

v1
sin() cos() r
vx
v2 sin()
cos() r
=
vy
v3 sin() cos() r

v4
sin() cos() r
Robots wheels are specially designed so that they do not

cause so much trouble while moving sideways. Each wheel

has 15 rollers and this is pretty much enough to realize
smooth movements.
This framework is intended to be as modular as possible
so that newly developed path planning algorithms can be
inserted or replaced with the existing ones with almost no
effort whatsoever.
As it is shown in figure 4, there are multiple processes
that are dedicated to different tasks such as providing global
information, deciding on the goal positions, planning paths
for a given current-goal configuration pair, and transmitting
action commands to the robots. This modular architecture
helps us to replace simulated environment with the real
robotic system without changing the algorithmic content of
the system.
Below are the descriptions of the nodes in the ROS
network.
ssl sim vision, Small-size-league(SSL) simulated vi-

Fig. 4.

This graph is auto-generated by ROS rxgraph utility. It shows how the messages are transmitted through which node(=process)

sion module. It publishes global pose of the robots that

are inside and just-outside of the field with proper state
flags (inside-field, outside-field, out-of-fov etc.) to the
global state topic. It is also a supervisor node in the
webots environment.
ssl game planner, SSL game planner. Currently it publishes predefined target poses for all robots, but it is
planned to be a hybrid/deliberative robot controller in
which there might be several modules -nodes- later on.
It subscribes to the global state topic and publishes the
target poses of the robots to the team pose command
topic.
ssl path planner, SSL path planner. In this study, this
node runs attractive/repulsive potential fields or navigation potential function based path planners. It subscribes
to the team pose command and global state topics, and
publishes to the visualization and team command topics.
ssl sim transceiver, SSL simulated transceiver. It is
a bridge between the robots and offboard computing
units. Subscribes to the team command topic and emits
messages to the simulated robots through webots communication interface.
ssl sim visualizer, SSL simulator visualizer. It helps to
visualize the path taken by the robots and the commands
given to them.

Since there is not a theoretical way to select best , selection

process may require serious intuition and effort.

Fig. 5. Left: additive att/rep potential function, right: navigation potential

function.

B. Case: Three-obstacle indirect-free path

IV. DISCUSSION
The two potential functions mentioned in this report,
namely additive attractive/repulsive potential function and
navigation potential function are tested in three different scenarios, one-obstacle direct-free path, three-obstacle indirectfree path, two-obstacle direct-free path. Please see the section
VI for the links to the videos representing each experiment
case.

Fig. 6. Left: additive att/rep potential function, right: navigation potential

function.

A. Case: One-obstacle direct-free path

In this experiment, we would like to see if the robot
can traverse around a single obstacle to reach the goal
position by following a preferably linear path. Normally,
since the obtacle is not entirely in the path of the robot, we
would expect that the robot almost-directly goes to the target
position. But it is shown that without doing a visibility check
basic additive potential fields may cause perturbances in the
path or even oscillations depending on how the attraction
and repulsive gains are set. Navigation functions, on the
other hand, may give a better path if is properly chosen.

As it is shown in figure 8, basic att/rep potential function

method suffers from the local minima problem and robot gets
stuck at some point. On the other hand, navigation function
finds a path to the goal, though far from being optimal.
Figure 7 shows how the path length varies with different
values.
C. Case: Two-obstacle direct-free path
Although this case is similar to first case, it shows the
severity of the additive att/rep potential fields which even

TABLE I
OVERALL RESULTS
Experiment
Case 1
Case 2
Case 3

Additive
time(sec.) path(m)
2.5
1.25

Navigation
time(sec.)
path(m)
2.5
1.26
45.52
1.97
1.5
1.09

To deal with local minimum problem in additive functions

one idea would be applying random forces on the robot once
it is detected that the robot stops without reaching to the
goal position (viz. getting stuck at a local minimum). Similar
method is proposed in [3], called random walk.
Fig. 7. From top left to the bottom right values are 3,4,5 and 7. The
time it takes to reach to the goal is t3 =no value, t4 = 62s, t5 = 53s
and t7 = 45s in simulation time. Thats why is chosen to be 7 in the
experiments. Please notice that the viewpoints of these figures are different
than each other, careful inspection will reveal the fact that as values
decrease, the length of the path increases since the shape of the bowl=like
potential function becomes more spreaded-out.

may not result a complete solution, although the only thing

that the robot should do is going just forward.
Navigation function, on the other hand, could find a path
between the obstacles, though the movement was a little
oscillatory. One reason of the oscillation is the delay between
the communication nodes(=processes) since the PC that Ive
used for the experiments was a standard laptop.
The overall results for the three experiment cases are given
in the table I. Although additive potential functions are lack
of completeness, in most of the cases, they provide shorter
paths than the paths navigation functions return.

Another solution to the local minima problem would be

using repulsive forces in a different form. For instance, if the
robot is imagined as a fluid or gas particle, at least in the
sphere world, robot will be able to pass through obstacles
standing between the robot and the goal position. However,
this method will not be successful in the case of confronting
narrow passages, causing collision to the obstacles. Another
problem this method will cause is the oscillatory movements
in locally cluttered environments in which there might be
paths that are longer in the sense of euclidean distance
but shorter in time since robot can reach higher speeds in
obstacle-free paths.
One solution to the oscillatory movement problem would
be grouping the obstacles together and having virtual obstacles which may contain narrow passages, and the obstacles
that are sufficiently close to each other as it is shown in
figure 9 and 10. This method with the visibility testing is
proposed in [2], and I have successfully implemented this
method(given in the algorithm 1), and tested in the simulation
environment in one of my earlier studies. Although tests are
done in highly cluttered environments, generated paths were
close to the optimal paths and no local minima problem is
observed.
Algorithm 1.

Fig. 8. Top: additive att/rep potential function, bottom: navigation potential

function.

Obstacle Grouping

while preObstList.size() > 0 do

closestObst getClosestObst()
if isVisible(closestObst)=TRUE then
tempObst closestObst
repeat
if
isLinkable(tempObst,preObstList[i])=TRUE
then
linkObsts(tempObst, preObstList[i])
tempObst tempObstNeighbor
end if
until preObstList.size()= 0
end if
virtualObst closestObst
groupObsts(virtualObst)
postList.push back(virtualObst)
end while

Fig. 9. Closest obstacle and its neighbor obstacle are linked to each other
and nwe virtual obstacle makes obstacle C invisible to the robot (adapted
from[?])

Case: Three-obstacle indirect-free path

Additive attractive/repulsive potential function
http://www.youtube.com/watch?v=
XreCjhkEJA8.
Navigation potential function http://www.
youtube.com/watch?v=AaW1E8bh9kg
Case: Two-obstacle direct-free path
Additive attractive/repulsive potential function
http://www.youtube.com/watch?v=
hGoSAFj7SL4.
Navigation potential function http://www.
youtube.com/watch?v=k8oT6fFCZaM
Obstacle
grouping
algorithm
http://www.
youtube.com/watch?v=0rp-VhkjYd0
R EFERENCES

Fig. 10. Obstacle 3, 4 and 5 are linked to each other.They form a virtual
obstacle (adapted from [2])

V. CONCLUSIONS AND FUTURE WORK

The main focus of this study was to compare simple additive attractive/repulsive potential function with the navigation
potential functions(for various values). The results show
that it is difficult to find a proper value for navigation
potential functions, and simple additive attractive/repulsive
potential functions suffer from the problem of getting stuck
in local minima. In order to use these methods in a real-life
application, like a robocup small size league robot soccer
domain, it is better to utilize other functionalities such as
visibility and path-free checks so that irrelevant obstacles
can be discarded while calculating the potentials. It is shown
that grouping the obstacles might result in shorter paths
considering the time it takes to the goal position and it
provides a better representation of the environment rather
than handling every obstacle seperately.
VI. ACKNOWLEDGEMENT
All the source code; ssl sim vision, ssl game planner,
ssl path planner, ssl sim visualization, ssl sim transceiver
can be found in this repository.
This document could also be downloaded in pdf format,here
Here is the list of videos:
Case: One-obstacle direct-free path
Additive attractive/repulsive potential function
http://www.youtube.com/watch?v=
ra-4UXbUaz0.
Navigation potential function http://www.
youtube.com/watch?v=QWG0X2xIJ-Y

[1] Principles of Robot Motion by Howie Choset etal., MIT Press, 2005,
ISBN: 0-262-03327-5
[2] B. Zhang,W.Chen, M. Fei, An optimized method for path planning
based on artificial potential field, Sixth International Conference on
Intelligent Systems Design and Applications (ISDA06) Volume 3
(2006)
[3] J. Barraquand and J.-C. Latombe. Robot motion planning: A
distributed representation approach. Internat. J. Robot. Res.,
10(6):628649, 1991
[4] Quigley M., Conley K., Gerkey B., Faust J., Foote T. B., Leibs J.,
Wheeler R., and Ng A. Y. (2009). Ros: an open-source robot operating
system. n International Conference on Robotics and Automation, ser.
Open-Source Software workshop.
[5] Michel, O. Webots: Professional Mobile Robot Simulation, International Journal of Advanced Robotic Systems, Vol. 1, Num. 1, pages
39-42, 2004
[6] R. Jarvis. Coliision free trajectory planning using distance transforms.Mech. Eng. Trans. of the IE Aus, 1985.
[7] J. Barraquand, B. Langlois, and J.C.Latombe. Numerical potential
field techniques for robot planning.IEEE Transactions on Man and
Cybernetics, 22(2):224-241, Mar/Apr 1992.
[8] Koditschek, Daniel E., and Rimon, Elon: Robot navigation functions
on manifolds with boundary, Advances in Applied Mathematics 11(4),
volume 11, 412442, 1990.
[9] Rimon, Elon and Koditschek, Daniele; Exact robot navigation using
artificial potential functions IEEE Transactions on Robotics and Automation. Vol. 8, no. 5, pp. 501-518. Oct. 1992