Analysis of safe motorcycle-following distance to prevent rear-end collisions | |
| Author | Phanuphong Prajongkha |
| Call Number | AIT Diss. no.TE-23-02 |
| Subject(s) | Rear-end collisions--Thailand--Prevention Motor vehicle driving--Thailand--Safety measures Traffic safety--Thailand--Mathematical statistics |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Transportation Engineering |
| Publisher | Asian Institute of Technology |
| Abstract | The purpose of this research was to develop a motorcycle (MC)-following protocol to address an issue concerning MC rear-end collisions in Thailand. There are three primary objectives of this study: The first aimed to identify a safe following distance (SFD) concept. It involved comparing SFD with conventional models and VDO footage taken from actual MC rear-end collisions in order to validate the outcomes. The second objective analyzed factors contributing to critical following distances. The third objective was to estimate the fatality probabilities among MC riders involved in rear end crashes. This dissertation was structured into three distinct studies, the outcomes of which are summarized as follows:Objective 1 The first part of this study aimed to categorize MCs’ distance-following using trajectory traffic data and validated the results with real MC rear-end collisions. Analysis was done using a dataset containing 8,223 events of MCs following a leading vehicle in Pathum Thani, and 41 cases of real MC rear-end collisions in Thailand from 2017 to 2021. The proposed SFD concept incorporated time headway (TH), safe stopping distance (SSD), and time to collision (TTC) to classify MC-following (MF) situations. The SFD concept is composed of three lines: 1. Longitudinal initial distance-following (Ln): The initial distancing from the MC to its leading vehicle, which affects the acceleration and deceleration behavior of the following MC. 2. Longitudinal warning distance-following (Lw): This indicates that the following MC can avoid immediate rear-end collisions from its leading vehicle's sudden braking. Still, there is a potential for a rear-end collision due to other factors such as slippery roads, braking failures, or loss of control. 3. Longitudinal critical distance-following (Lc): This shows that the following MC has limited opportunity to prevent an accident due to insufficient following distance if the preceding vehicle brakes abruptly.The first objective’s results showed that the SFD concept can identify potential rear end collisions. To validate whether the SFD concept could explain the risk of rear-end crashes, VDO clips of real MC hitting the rear-end of its leading vehicle were investigated. The following distance of the MC and its speed were used to validate the SFD concept. The results showed the plots of speed and following distance of 41 rear end crashes extracted from the clips. Notably, the majority of MC rear-end collisions (38 cases, 92.68%) occurred within the Lc area, indicating that there may not have been enough time for evasive maneuvers due to the limited following distance. Furthermore, Lc showed the risk of rear-end collisions, which accounted for 92.68% of the total cases. However, only three incidents of rear-end collisions in the Lw area were identified, which is very few. The findings highlighted the significance of the SFD concept, offering a possibility for the prevention of MC rear-end collisions by implementing reminder systems that notify riders when approaching the warning and critical areas of following distance.Objective 2 The second part was to analyzes the influential factors affecting MCs’ distance following within the Lc. In this part, 8,223 observations were employed. The TTC value, which was developed from objective 1, was used to build the Lc concept. Binary logistics regression analysis was applied to investigate the relationship between the independent variables and the dependent variable, which was assumed to be a dummy variable. If the actual following distance is smaller than Lc, it would be assumed to be 1 in a critical situation; otherwise, 0. From these conditions, logistics regression can be used to estimate the probability of Lc between 0 and 1 by the logistic distribution function. The findings indicated that high-speed usage significantly influences the likelihood of being within the Lc area. The relative speed displayed a considerable positive effect, with a p-value of 0.000. This suggests that if riders increase their speed, thereby affecting their relative speed, they are likely to follow a vehicle ahead closely with a similar speed to the leading vehicle. If helmet usage was considered, riders who wore helmets showed a negative effect. This implies that riders who wear helmets are more inclined to maintain a larger following gap compared to those who do not. Helmet usage appeared to reflect a more cautious riding behavior, as evidenced by a larger following gap from the leading vehicle. Riders without helmets exhibited a higher probability of approaching Lc. Notably, in terms of a single-leading vehicle, the model displayed significant positive effects. This indicates that riders following a single leading vehicle tend to keep a smaller gap than those following multiple vehicles. Concerning the factor of leading vehicles (based on leading cars), both categories exhibited a significant negative effect. This suggests that riders following a MC, or a truck tend to maintain a greater gap compared to those trailing a car.Objective 3 The third study examined the probability of MC riders’ deaths when their vehicle rear ends a leading vehicle in Thailand. From 41 real MC rear-end collisions, thirty cases were injured, eight were fatal, and the remaining three indicated that the cause of fatality was due to double crashes with other vehicles. Based on the VDO evidence, the results indicated that the largest proportion of the MC rear-end collisions, accounting for 82.9%, occurred without a pillion passenger, whereas just 17.1% showed a pillion passenger. In terms of safety equipment usage, 73.2% of riders wore helmets, while 26.8% did not. When collision avoidance strategies were analyzed, 29 cases (70.7%) were found to have used just brakes. Additionally, a minority of strategies, only two cases (4.9%), demonstrated a combination of evasive actions and braking. Interestingly, some riders (8 cases, 19.5%) did not execute any action to avert the collision. A deeper analysis revealed that 7 out of 11 fatal cases, were emergency reactions involving braking without evasive actions that were prone to resulting in fatalities. This part also employed binary logistic regression between fatality and non-fatality outcomes among MC riders. The analysis was divided into two parts: the first part analyzed the 41 real MC rear-end collisions, and the second part analyzed 38 real MC rear-end collisions that occurred within Lc area. According to 41 real MC rear-end collisions, both pre-impact speed and relative speed exhibited statistical significance. Both pre-impact speed and relative speed showed positive signs, suggesting that riding at higher speeds and experiencing higher relative speeds would increase the probability of mortality in MC rear-end collisions. The developed models in this study indicate that the 50th percentile of fatality probability is approximately 98.6 and 54.44 km/h for pre-impact speed and relative speed,respectively. These results help fill a significant information gap regarding the risk of MC rear-end crashes and the potential factors contributing to fatalities. Based on 38 real MC rear-end collisions, the pre-impact speed significantly increased the fatal probability of the following vehicle. The model of the present study demonstrated that 50% of the fatality probability from models is approximately 26.86 m/s (96 km/h) for pre-impact speed. Riding at high speed and high relative speed might lead to fatal injury. It is important to note that not all fatalities may be avoided by wearing a helmet; in fact, this study found four fatal cases involved the use of helmet. This implies that the fatality result is caused not only by an impacted head, but also by other injured body parts such as the neck, thorax, and entrails, which can contribute to the fatality result due to the severe force impaction from high-speed use. Also, the findings indicated that the probability of fatality due to MC rear-end crashes significantly increases when riding at high speeds and when the MC hits the rear-end of a truck. Therefore, speed management in specific MC must be strictly enforced.Summary According to the 1st, 2nd, and 3rd objectives, the findings greatly contribute to comprehension of safe following gaps, particularly for MCs, and the study of the influential factors leading to risks, including the potential for fatalities resulting from rear-end collisions. These results have great potential to assist transportation safety authorities and MC manufacturers by providing insights into appropriate distance following practices and the causes of MC rear-end collisions. Such insights can, in turn, facilitate the development of sustainable safety reminder systems. Furthermore, by employing these variable insights and advocating for regulations that prioritize safe following gaps, these findings can enable stakeholders to concentrate on promoting MC safety. The findings, in this context, are poised to support transport safety authorities and MC manufacturers in understanding the risks faced by MC riders. Therefore, this will encourage the advancement of road safety in Thailand. |
| Year | 2024 |
| Type | Dissertation |
| School | School of Engineering and Technology |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Transportation Engineering (TE) |
| Chairperson(s) | Kunnawee Kanitpong |
| Examination Committee(s) | Santoso, Djoen San;Surachet Pravinvongvuth |
| Scholarship Donor(s) | Royal Thai Government;AIT Fellowship |
| Degree | Thesis (Ph. D.) - Asian Institute of Technology, 2024 |