Impact of Road Geometry and Land Use on Motorcyclist Driving Stress

Road rage is crucial in understanding the relationship between driving stress, risk inclination, and traffic sanctions [1]. Although generally overlooked by drivers [1, 2], asserted that it predicts hazardous driving behavior. Tension can negatively affect driver performance, increasing the likelihood of traffic violations and the risk of road accidents [3-5]. Previous research highlighted that driving, a significant part of commuting, can be a complex and stress-inducing activity, often caused by interactions with other drivers [6, 7]. Moreover, prolonged driving can induce a continuous stress response [8]. Previous research has explored the influence of stress on driving performance, behavior, and road safety [9, 10].

In Indonesia, the accelerating pace of transportation and inadequate infrastructure contribute to congestion, longer driving times, and increased fatigue, which raise accident risks [11]. Continuous driving leads to fatigue and lethargy, increasing the likelihood of human error and accidents [12]. Aggressive driving, linked to road rage, significantly raises accident risks. Drivers who frequently violate traffic rules and understand speed limits tend to exhibit more aggressive behavior [13, 14]. Studies have identified hostile aggression associated with rage [15], and a meta-analysis demonstrates a strong correlation between driving rage and accidents [16]. Furthermore, drivers often exhibit anger as a coping mechanism in response to aggressive behavior from others. Research shows an association between anxiety and aggressive driving, with 80% of drivers reporting some anxiety, particularly among women over 35, where anxiety levels are higher than in men [17-19].

While problem-focused coping strategies were more commonly employed by drivers exhibiting instrumental rather than hostile aggression [20], numerous transportation challenges, such as traffic congestion, can significantly intensify driving stress. When the volume of vehicles surpasses the capacity of the road, journey times are adversely affected, thereby exacerbating stress levels and increasing the likelihood of accidents. As stipulated by Government Regulation No. 44 of 1993, article 240, paragraph 2, concerning Vehicles and Drivers, the maximum daily driving time is set at eight hours; exceeding this limit further exacerbates tension levels [21] further increases stress, while poor stress management impairs individuals' ability to interact with their environment [22].

In Indonesia, particularly Banda Aceh, high motorcycle dependency has led to greater accident exposure in urban areas [23]. As a growing capital city, Banda Aceh has experienced rapid development and congestion, contributing to 72% of traffic [23], driven by economic growth and supported by government investment in education [24]. With a population of 265,111 and a density of 43 people per, traffic pressure continues to rise. Motorcycles make up 80% of motorized traffic [25], with registrations increasing from 219,532 in 2020 to 250,154 in 2024, representing 13,95% of all vehicles [25-28]. This growth aligns with rising accidents involving motorcycles, pedestrians, and bicycles [29]. In 2020, there were 103,228 traffic incidents and 30,568 deaths in Indonesia, mirroring the global trend of road accidents being the leading cause of death among those aged 14 to 29 [30].

In Indonesia, three people die every hour due to road accidents, with 61% caused by human factors, 9% by vehicle issues, and 30% by infrastructure and environmental conditions [31]. Drivers under work-related stress tend to exhibit risky behavior, while those without stress are likelier to follow occupational safety practices [31]. Studies show men are more prone to unsafe driving than women, though experience reduces such tendencies [32]. The diverse traffic mix—cars, motorcycles, rickshaws, and more—adds complexity to road safety [33]. In Banda Aceh, traffic grows by 6% annually, driven by rising living standards and mobility demands [34]. Urban transport issues stem from uneven road networks, activity concentration, and a shift to private transport like motorcycles, prompting infrastructure expansion to meet growing regional mobility needs [35]. Additionally, extreme stress can trigger physiological responses such as increased heart and respiratory rates, dilated pupils, muscle contractions, and anxiety, further affecting driver behavior [36].

Previous research has shown that driving tension directly or indirectly contributes to traffic accidents among professional drivers [37], with a strong correlation between drivers’ physiological responses and stress levels [38]. Meilinda [39] found that driving stress is associated with road geometry, particularly on curved roads and in office or industrial areas. Road design and urban spatial structure significantly influence how drivers manage vehicle movement and stress levels. In Banda Aceh City, changes in land use due to urban development indirectly affect traffic distribution patterns. Therefore, this study aims to examine the factors influencing motorcycle driving stress in Banda Aceh by analyzing road geometry and land use patterns. Data were collected using heart rate variability (HRV) and analyzed through a Multiple Indicators Multiple Causes (MIMIC) model.

3.1 Study area and data collection

The study was conducted along the primary arterial road, specifically Jl. T. Nyak Arief and Jl. T. Muhammad Daud Bereueh. The road was divided into thirty 500-meter segments for each respondent. Figure 1 depicts the location of the primary arterial road in relation to this research. Over a period of approximately three months, from June 26 to August 10, 2022, a psychological indicators survey instrument was compiled for several respondents who were selected as research samples. Data collection sessions were scheduled during peak hours in the morning, afternoon, and evening, as well as on weekdays and weekends, based on the availability and approval of the respondents.

1.png

Figure 1. Area of the study (map created by the authors)

2.png

Figure 2. (a) Vantage V2 polar watch [59]; (b) GoPro MAX 360 [60]

The instruments utilized in this research comprised a Vantage V2 polar watch and a GoPro Max 360 video camera, as depicted in Figure 2. The Vantage V2 polar watch was employed to acquire GPS data and featured a heart rate sensor. The participants donned polar watches equipped with close proximity sensors to guarantee accurate detection. These sensors then transmitted heartbeat signals to a receiving device operating the polar flow application. Conversely, the GoPro Max 360 video camera captured footage of various road conditions, such as highways, lanes, traffic lights, bridges, medians, and overall traffic conditions. Moreover, the sensor measured the heart rate of the subject 10 minutes prior to driving, serving as the primary condition or "rest condition," and while traveling along a predetermined route. The sensor then transmitted the heart rate signal to the Polar Vantage V2 watch, which was extracted using the Flow app. A total of 25 respondents were selected based on specific criteria, which included factors such as being in good health, holding a valid motorcycle driving permit, and a balanced representation of students and workers, as well as an equal number of males and females. Respondents were randomly selected to ensure diversity, and their willingness to participate in the research study was confirmed. Based on the outcome of the data evaluation, 22 datasets were collected, considering factors such as land use, road geometric and traffic control, from 25 respondents. Five hundred fifty (550) datasets were obtained for the primary arterial road.

3.2 Multivariate analysis with MIMIC model

The multiple-indicators multiple-causes (MIMIC) model is commonly employed in transportation research focused on travel behavior. This model is used to thoroughly examine individual behavior and the psychological perceptions of respondents [61]. In a study conducted in Jakarta, the SEM-MIMIC model was utilized to investigate the relationship between public perceptions of transportation system policies. Additionally, an in-depth analysis of public acceptance of transportation policies in Jakarta was conducted using the multiple-samples multiple indicators multiple-causes (MS-MIMIC) model [62, 63]. The objective of this analysis was to demonstrate the connection between travel behavior and public perception of transportation policy. The MIMIC model consists of structural equations and measurement models, as outlined in Eqs. (1) and (2).

$\eta_{\mathrm{i}}=\mathrm{B} \eta_{\mathrm{i}}+\Gamma \mathrm{z}_{\mathrm{i}}+\zeta_{\mathrm{i}}$                (1)

${\mathrm{y}_\mathrm{i}}=\Lambda \eta_{\mathrm{i}}+\zeta_{\mathrm{i}}$                  (2)

where, $y_i$ is a vector of perception indicators (unobserved variables), $z_i$ is a vector of observed variables $\eta_i, \, B, \, \Gamma$, and $\Lambda$ matrices of regression coefficients that must be estimated (unknown parameters), and $\zeta_i$ and $\zeta_i$ are vectors of measurement error. The Maximum Likelihood Estimator (MLE) method is implemented to determine unknown parameters, and its programming is executed using the SIMPLIS common language, as specified in reference [64]. Furthermore, the program is executed using LISREL 10.20 software. The MIMIC model incorporates unseen latent variables, which are assessed through various indicators and predicted by different causes. By conducting a thorough examination of each score of the research variables, the structural equation modeling (SEM) data analysis enables a comprehensive evaluation of the constructs. The questionnaires or statements used in the study are considered as either manifest or latent indicators of the variables being measured [65].

The MIMIC model establishes the connections between observable and unobservable variables by minimizing the distance between the sample covariance and the model-predicted covariance matrix. SEM analysis is particularly effective in addressing non-experimental research problems. The general SEM model comprises two components: the measurement component, which links the tested variables to the latent variables through a confirmatory factor model, and the structural component, which connects the latent variables through a set of equations simultaneously [65, 66]. In addition, it is common practice in the field of SEM to evaluate the validity and reliability of measurement instruments through confirmatory factor analysis [66, 67]. Specifically, this analysis is conducted in two stages: exploratory factor analysis and confirmatory factor analysis. Exploratory factor analysis is used to identify potential factors in the data, while confirmatory factor analysis is used to test and refine theoretical models over time [67, 68].

3.3 Data analysis method

The HRV parameters were preliminarily analyzed using Kubios HRV 3.5.0 software, followed by further analysis and calibration via the MIMIC method. This method employs the Maximum Likelihood Estimator in LISREL 9.2 software to estimate the parameters. The MIMIC model is designed to test the null hypothesis of the model and determine the vectors of indicator variables (y) linked by a latent variable ($\eta$) and a covariate (x). The model comprises two equations, one of which examines the relationship between (x) and ($\eta$) and the other confirms the relationship between (y) and ($\eta$). To evaluate the model's viability and the parameters' accuracy, Goodness of Fit (GoF) model testing will be performed. The GoF assessment will involve several metrics, including GFI (> 90%), AGFI (> 90%), CFI (> 90%), RMSEA (≤ 0.08), and T test for parameter significance <5% [61, 62, 69, 70].

Based on the reference range of 60 to 100 beats per minute in Table 2, which represents the normal adult heart rate, the derived heart rate (HR) value suggests that the driver's overall condition is approaching the stress threshold. The term bradycardia refers to an irregular heartbeat where the heart's rhythm drops to or below 60 beats per minute [73]. Similarly, tachycardia is a condition characterized by a heart rate exceeding 100 beats per minute [53]. It is worth noting that urban areas with geometric curves or bends tend to elicit the highest heart rates. Interestingly, the research findings indicate that the threshold at which an individual's heart rate falls below the normal limit is only 6%.

This research highlights the relationship between driving stress due to the influence of differences in land use and road geometry in urban arterial road areas using the heart rate variability (HRV) method with two approaches, namely stress time domain and stress frequency domain. The evaluation and interpretation of heart rate fluctuations is the primary distinction between the HRV analysis's stress time and frequency domains. The stress time domain focuses more on the general fluctuation of heart rate over a certain period, while the stress frequency domain focuses more on the distribution of heart rate activity in various frequency ranges related to autonomic nervous activity. These two approaches can provide insights into drivers' physiological reactions of drivers to different driving conditions and traffic patterns.

The empirical results showed the tendency distribution of driver characteristic data, with the stress time domain significantly affected by the type of office land use. Stress based on the heart rate cadence of the driver had a greater impact on inanimate objects, such as infrastructure and land use, than on the driver. Tension from the nervous system tended to significantly affect objects in direct contact with the driver, such as geometric conditions of road bends, ascents, and descents. Increased stress was attributed to high traffic congestion, particularly evident during rush hour when the observations were conducted. Therefore, it was assumed that office areas also experienced heightened traffic congestion. Mixed traffic congestion in these areas during peak hours further increased stress levels in drivers.

This study demonstrates how land use characteristics, like traffic density, the kind of driving behaviour, and the placement of buildings or other infrastructure, can influence the level of activity and distractions experienced by drivers. For instance, a high traffic density near offices might lead to higher levels of stress and tension, which can be measured by a faster heart rate. Stress from driving is highly impacted by the office land use variable (standardized coefficient=0.70, p < 0.01). Because of the rush hour, loud surroundings, and concentration of cars, hectic office spaces might make drivers more stressed. The motorist may experience an increase in heart rate in response to stress due to all of these factors affecting their cardiac rhythm. High traffic and congested locations, such industrial districts (non-standardized coefficient=0.39, p=0.05, standardized coefficient=0.12), might make drivers feel more stressed out. Heavy traffic, truck, and mixed-road uses also have a positive and significant impact on driving stress [38]. Heavy traffic is frequently seen in industrial and office locations, especially during rush hour when workers are commuting to and from work. Drivers may experience tension and anxiety due to this dense traffic.

Additionally, driving stress is highly impacted by the educational land use variable (standardized coefficient=0.49, p=0.002). The driver's route included an elementary school and a state institution. The findings indicate that drivers may experience stress when operating a vehicle in an educational setting because of the requirement to protect children's safety and security, particularly in the vicinity of primary schools. Stressful situations for drivers include having to watch out for pedestrians, potentially careless kids, and unique traffic patterns around schools. Areas around schools also often have limited parking, which can make drivers spend more time relocating their vehicles. Furthermore, drivers may feel more stressed because of the busy and boisterous environment surrounding schools, particularly at the start and end of the day. These factors were obtained from direct observations in the field after being carried out several times at the research location.

However, some elements can give the driver a direct physical stimulus, like turn in the road (standardized coefficient=0.48, p=0.01), turns with derivative and incline (standardized coefficient=0.60, p=0.01), and incline conditions and derivatives (standardized coefficient=0.82, p=0.01). Road conditions, steep inclines, and sharp turns can all make it more difficult for drivers to maintain focus, control, and motor reaction, which in turn can make them feel more stressed. Nervous system-related stress reactions, such heightened attention, tense muscles, or elevated adrenaline, are more closely linked to the physical state and perception of the driving environment by the driver. Road geometry conditions that are hazardous or physically taxing can trigger this physiological stress response.

Driving around a bend can be challenging for drivers because of derivation and inclination bends. This may be the result of the slope decreasing line of sight and obstructing the view. Additionally, drivers need to be able to anticipate how the car would react to variations in road height. Similar results were shown from research [36], which stated that segments sharp turns can cause driving stress because the driver concentrates on the sharp turns and vehicles from the opposite direction. On the other hand, the fact that sharp turns are places where dangerous maneuvers cause driving stress is in line with previous research [74].

Land use conditions therefore have a substantial impact on stress based on heart rhythm since they affect the driver's degree of activity and distractions. On the other hand, as they affect the driver's physical and physiological sensations, road geometry can have a substantial impact on stress that is based on the nervous system. The combination of these two elements creates a convoluted driving environment and influence drivers' stress responses around areas with heavy traffic. Finally, it should be noted that locations with mixed and confused traffic, lane changes, road crossings, sharp turns, and traffic jams all cause the most driving stress.

This research also emphasizes the importance of the interaction between land use conditions and road geometry in influencing driver stress levels. The practical implications of these findings are highly relevant for urban planning, traffic safety, and transportation policy. Identifying areas with a high potential for driving stress such as office, industrial, and educational zones can support the development of more targeted interventions to reduce the psychophysiological burden on motorcycle riders. Strategic measures may include installing speed-calming devices (e.g., roundabouts or speed bumps), improved signage, and better road design in high-risk areas such as sharp curves, steep inclines, and congested urban corridors. A deeper understanding of how environmental factors contribute to driver stress can facilitate the creation of safer, more adaptive, and human-centered traffic systems.

Moreover, the study recommends integrating physiological data from drivers such as heart rate variability monitored through wearable devices into smart urban traffic management systems. This approach enables dynamic identification of high-risk road segments and can inform responsive safety measures. In the long term, urban zoning policies may consider the health implications of stress exposure, especially for motorcyclists, and support spatial separation between high-activity zones (e.g., offices and industrial areas). The study also contributes to developing Intelligent Transportation Systems (ITS), particularly for two-wheeled vehicles, which have often been overlooked in such innovations. Integrating biometric data into vehicle design or navigation systems could enhance driving safety by providing real-time alerts in high-stress areas or suggesting alternative routes. Therefore, the outcomes of this research not only enrich the academic discourse but also offer actionable insights for policymakers and urban planners to establish safer and more responsive urban transportation networks.

The authors affirm that they do not possess any identifiable personal relationships or competing financial interests that might have appeared to exert an influence on the research presented in this article.