Effect of Response Test on Car-Driver Alertness
Traffic accident in Indonesia seems getting worse for the last few years. It was increasing significantly, even for the comparison of the number of traffic accident to number of vehicles operated in Indonesia. Data shows that during 2004-2008, number of accident that involved car-vehicle increased for about 31.43%, while accident that involved bus and truck increased 28.29% and 34.60% respectively. This is supposed to warn every related institution to give significance contribution in order to minimize the traffic accident in Indonesia. This research was conducted to be one of them.
One of the most dangerous behavior or condition that lead driver to a traffic accident is lack of alertness. Sleep deprivation and physically or mentally fatigue condition could both lead to impairment of alertness while driving. Other condition such as monotony, distraction, and drug/alcohol could also reduce drivers’ alertness level (Desai and Haque, 2006). It is clear enough that to prevent traffic accident, we have to maintain drivers’ alertness, because driving activity requires the driver to maintain high level of alertness.
Previous researches in Cognitive Ergonomics showed that auditory activity closely related to alertness. The most common behavior in driving activity is listening to the radio or increasing the volume when sleepy/fatigue. However, some studies showed that this not an effective solution, since driver could be more sleepy than before (Oron-Gilad and Shinar, 2000). Another type of auditory activity is warning signal or alarm, which are more effective to give alerting effect to the drivers.
Furthermore, the next issue of an additional auditory activity is that whether it can interfere with the primary task itself. On the driving system, driving the vehicle is the primary task. Therefore, the challenge in this study is to design auditory activity as a secondary task that able to maintain drivers’ alertness without any distraction on the driving task.
This study proposed a warning signal as the auditory activity, and referred as a response test. This response test was build manually by Visual Basic programming language and run in a set of personal computer, complete with the audio speaker and a push-button. While driving, driver would hear warning sign from the speaker and push the button.
A laboratory experiment was carried out in a driving simulator; build in Laboratory for Work System Design and Ergonomics, ITB. The hardware system used Logitech G-25 TM set of pedal, steering wheel, and six-transmission gearshift. A set of personal computer, audio system, and projector were also utilized. The software system represented Indonesian highway situations, built by 3D Game Studio®. A loop route of driving was applied, with urban and suburban areas on each side of the driving route, marked by skyscraper buildings and woods, consecutively.
Drivers’ alertness level was observed by four gold electrodes of BioGraph Infiniti electroencephalograph (EEG) (Fz, Cz, Pz, Oz (Jung, 1997; Berka et al., 2005)). Each of them was referenced to the ear lobe and sampled at 250 Hz. Computation was then conducted by Power Spectral Density (PSD) to gain theta (4-8 Hz), alpha (8-13 Hz), and beta (13-20 Hz) energy level (Fisch, 1991; Lal et al., 2003). Combination of the spectral power in these frequency bands were used to get computation of the α/β and (α+θ)/β indices, which were associated with decreasing of alertness (Oron-Gilad et al., 2008; Rogé et al., 2009).
Two types of response test method were tested: one with random time interval (among one, two, three, four, or five-minute interval) and another with three-minute interval of occurrence. Counterbalancing method was applied to minimize bias from the sequence of response test activity. This response test was played on 82 dB, while the environment was stable on 70 dB range of sound. Reaction time of each warning signal occurrence was also measured to give further analysis on drivers’ alertness. The increase in time taken to respond to the auditory stimulus (reaction time, RT), decrease in response speed (1/RT), and increase in lapses (responses > 500 ms) were computed as these parameter were associated with decreasing alertness (Baulk et al., 2008).
Eight male (aged 23.5 ± 2.33 years old) were involved as participants. However, only six of them finished the experiment. All of them had more than one year experience in driving car-vehicle. Before the experiment, they were filled out an informed consent form, containing the experiment procedure, participants’ rights and obligation, etc. Each participant drove the simulator for about one hour, consisted of half hour of random response test and another half hour of 3-minute response test.
Statistical analyses were then conducted by within-subject design of two-way ANOVA to find out the effect of response test type on alertness of several type of power spectrum indices. Result showed that there is no significant effect of the two types of response test on drivers’ alertness, indicated by calculation on the θ, α, β, α/β, and (α+θ)/β index (Fobserved =
0.005, Ftable [1/5] = 6.610). Statistical calculation on the reaction time was also brought similar result: no significant effect of the two type of response test on RT, 1/RT, and number of lapses (Fobserved = 0.222, Ftable [1/5] = 6.610). These results suggest that random response test has no difference effect on drivers’ alertness, compared with fixed-interval response test (3-minute interval for this study).
Furthermore, statistical computation was also conducted for analyze the effect of response test within the 3-minute interval type along 27 minutes period of the response test application. This time period was divided into nine segments; θ, α, β, α/β, and (α+θ)/β index were computed on each of these segments. Within-subject design of two-way ANOVA computation resulted in no significant difference among these time segments (Fobserved = 1.128, Ftable [8/40] = 2.180). This result showed that the proposed response test could maintain drivers’ alertness in the same level along the response test application.
Figure 1. Interface of proposed auditory response test
Figure 2. Documentation during the experiment.
LIST OF RESEARCH OUTPUT
1. Prototype of auditory response test apparatus.
2. Draft publication on:
a. Sutalaksana, I.Z., Mengko, T.R., Yassierli, Prihatmanto, A.S., Mahachandra, M. Aktivitas sekunder audio untuk menjaga kewaspadaan pengemudi mobil Indonesia. Jurnal Teknik Industri, Jurusan Teknik Industri, Fakultas Teknologi Industri, Universitas Kristen Petra Surabaya, ISSN: 1411-2485.
b. Sutalaksana, I.Z., Mengko, T.R., Yassierli, Prihatmanto, A.S., Mahachandra, M. (2011). Effect of auditory response test on Indonesian drivers’ alertness. IAENG International Conference on Industrial Engineering (ICINDE’11), Hong Kong, 16-18 March 2011.
c. Sutalaksana, I.Z., Mengko, T.R., Yassierli, Prihatmanto, A.S., Mahachandra, M. (2011). Analisis penggunaan waktu reaksi sebagai penentu tingkat kewaspadaan pengemudi mobil Indonesia. Prosiding Seminar Nasional Teknik Dan Manajemen Industri (SNTMI) 2011, Universitas Muhammadiyah Malang, 10 Januari 2011.
HEAD OF RESEARCH TEAM: Dr. Ir. Iftikar Z. Sutalaksana
TEAM MEMBERS: Prof. Dr. Ir. Tati R. Mengko, Yassierli, Ph.D., Dr. Ary Setijadi Prihatmanto, Manik Mahachandra, ST., MSc., Nur Indria Lisdiani, ST
OFFICIAL ADDRESS: Laboratory for Work System Design and Ergonomics, Industrial Management Research Group, Faculty of Industrial Technology, Labtek III building, 3rd floor, Ganesha 10, Bandung 40134,