INTRODUCTION
Few
researchers have addressed issues relating to the use of the joystick in ground
vehicle applications. In the choice of
design of future Army ground vehicles, and in future robotics and teleoperation
applications, drivers may potentially use joystick controllers rather than
steering wheels. Joystick control
provides several important benefits, including the advantage of utilizing the
flexibility and dexterity of the human hand, wrist and fingers to increase the
degree of driver control over vehicle operations. One researcher noted that if control and feedback parameters are
chosen carefully, a joystick controller may potentially outperform a steering
wheel with brake and accelerator pedals (Lee, 2000). The purpose of this study was to determine the effect of
different vehicle controllers on driver performance in a simulated tank driving
task. The results of this study will be
used to develop new U.S. Army crewstation concepts, and to provide more
knowledge for further studies such as the U.S. Army’s Crewstation Automation
Testbed, as well as Army Future Scout and Future Combat Systems.
METHODOLOGY
Subjects, Facilities and Apparatus
Subjects were
eight, U.S. Army Department of Defense male, right-handed civilian volunteers
solicited from the Human Research and Engineering Directorate (HRED) workforce
for the purpose of participating in this experiment. Subjects were screened for normal visual functioning. The experiment was conducted at HRED,
Building 459, Aberdeen Proving Ground, Maryland.
Apparatus
included a Silicon Graphics (SGI) Indigo Workstation. The steering wheel device was a Thrustmaster NASCAR Model Pro
Racing Steering wheel with pedal accelerator and brake. The joystick was a Logitech Wingman
Joystick.
Experimental Tasks
In the driving
task, the participant used the steering wheel or joystick device to drive on
digitized road terrain presented on the SGI Indigo workstation monitor located
directly in front of him. The
participant used the steering wheel accelerator and brake on the floor to
maintain speed, or manipulate the joystick device to accelerate, steer or brake
(pushing the joystick forward and backward to furnish acceleration and braking
action). The participant monitored his
speed by means of a speedometer shown on the workstation monitor. The
experimenter monitored the participant’s speed during the experiment to ensure
that the subject’s speed remained at or close to the required speed for that
session.
The terrain
was a digitized version of the Aberdeen Test Center Demo II Course at
Churchville, Maryland. Five different
courses were used, one for the training sessions, and one for each of the four
experimental sessions. The courses were
approximately the same length and had the same number of turns.
Differing
levels of driving task workload were manipulated by specifying different speeds
at which the participant performed the driving task. The speeds, which were representatives of low- and high-workload
tank driving conditions, were 15 and 45 mph.
Driving speed and control apparatus were counterbalanced across all
participants and experimental sessions to avoid practice effects.
Procedure
Prior to the
beginning of the experiment, each subject was given a half-hour training
session in which he was introduced to the driving task and the controller used
in the first experimental session. The subject then drove a practice test
course at 45 mph for 30 minutes.
After a
30-minute break, the subject performed the experimental driving task at the
pre-determined conditions of driving speed (15 or 45 mph) and vehicle
controller (joystick or steering wheel).
He was asked to maintain the assigned speed and remain on the course
while driving. When the driving task
was completed, the participant filled out a NASA TLX workload measure and had a
30-minute break. After the break, the
subject started the next experimental condition in which he drove at the
remaining pre-determined driving speed. When the driving task was completed,
the subject filled out a NASA TLX worked measure, and the subject was advised
to return the next day.
The next day,
the subject experienced one practice and two experimental sessions using a
different steering device, at the pre-assigned driving speeds of 15 and 45 mph.
When each driving task was completed, the subject again filled out a NASA TLX
workload measure. When both
experimental sessions were completed, the experiment ended.
Experimental Design
A 2 x 2 mixed
factor, repeated measures design was used for data collection and to structure
data analysis. Independent variables
were driving controller (joystick, or steering wheel with attached brake and
accelerator pedal), and assigned driving speed of 15 or 45 mph (the maximum
speed at which the subject was permitted to travel). Dependent variables were mean driving speed (the average speed at
which the subject actually drove), and the proportion of time the center of the
vehicle remained on the road during travel.
RESULTS
An Analysis of
Variance (ANOVA) was performed for each dependent variable to determine whether
statistically significant differences existed between main effects or
interaction effects in the study. Significant effects (p < 0.05) were
explored on a post hoc basis.
The ANOVA for
mean driving speed indicated significant main effects for controller (F
= 7.24, p = 0.031), for assigned driving speed (F = 2130.84, p
= 0.000, and for the control x speed interaction (F = 8.412, p =
0.023). The interpretation of the
higher-order interaction precludes the interpretation of the lower-order
interactions. The data for the
controller x speed interaction, which is plotted in Figure 1, indicated that at assigned speeds of 15 mph,
subjects obtained a mean driving speed of 14.7 mph using a steering wheel, and
a mean driving speed of 14.5 mph using a joystick. This difference was not statistically significant (p >
0.05). At assigned speeds of 45 mph,
subjects using a steering wheel obtained a mean driving speed of 39.9 mph, and
38.4 mph using a joystick. Although
this difference was statistically significant (p < 0.05), it may have
little practical significance because it is less than 5 mph.
The ANOVA
indicated no significant main effects or interactions for the proportion of
time the driver was able to keep the vehicle on the center of the road. Thus, vehicle controller made no significant
difference in the ability of subjects to maintain control of the vehicle.
Subject
comments provided insight into the use of the joystick controller. Although some subjects commented that the
steering wheel provided an advantage because of the large amount of prior
experience use in the automobile, others noted that the joystick seemed to
provide more control on curves. Some
drivers commented that the placement of the joystick was important, and that
arm support was needed during driving tasks in order to prevent driver fatigue.
CONCLUSIONS
The results
indicate that the steering wheel provided no significant advantage over use of
the joystick. At a relatively slow speed of 15 mph there was no significant
difference in mean driving speed between joystick and controller. At faster driving speed of 45 mph, the
difference between joystick and controller was small enough to have no
practical significance. In addition,
controller type played no significant role in assisting drivers to keep their
vehicle in the center of the road.
However, joystick controls have potential as an alternative control
technology because they provide more control over curves, possibly because they
permit greater use of the hand, wrist and fingers to increase the degree of
driver control.
Subject
comments implied that the ergonomic placement of the joystick could be an
important factor in enhancing driver performance. Subjects related that proper support of the hand, wrist and
forearm are important to efficient and effective joystick use. Future research should explore proper
joystick placement in the crewstation to allow optimum use in steering tasks.
REFERENCES
Lee,
J.D. (2000). A summary of issues concerning joystick control of ground
vehicles. Technical report prepared for Booz-Allen & Hamilton Inc. Iowa
City:
Author.
