- Can guide the driving experience via addl. mechanisms operated within motorbikes
- Local study proposes a motorbike that is developed using AI applications
Supervised artificial intelligence (AI)-based technologies could detect human factors that may contribute to breaking road regulations and poor judgement in advance, and convert the driving experience under the guidance of AI while poor environmental and road conditions can be addressed with additional mechanisms to be operated within the motorbikes by the AI.
These observations were made in an original article on ‘Road traffic accidents in Sri Lanka: A retrospective analysis and AI-based solutions for prevention’ which was authored by W.G.S.R. Thilakarathna, M.T.K.L. Thudugala, W.A.C.J. Hangilipola, W.N.S. Perera and P. Paranitharan (all five attached to the Kelaniya University's Medical Faculty's Forensic Medicine Department), and published in the Sri Lanka Journal of Forensic Medicine, Science and Law's 16th Volume's First Issue, last month.
Road traffic accidents continue to play a pivotal role in causing numerous fatalities, injuries, and socioeconomic challenges. Every year, approximately 1.19 million people are cut short as a result of a road traffic crash (the World Health Organisation's [WHO] ‘Global status report on road safety 2023’). In Sri Lanka, road traffic accidents reveal an upward momentum over the past decade.
According to the statistics for the year 2022 available with the Ministry of Transport, Highways, Ports, and Civil Aviation, the number of persons killed in road accidents involved 820 motorcyclists, followed by 720 pedestrians. Overall, 2,515 deaths took place in 2022. A total of 908 motorcycles were involved in fatal accidents, being the largest contributor (the National Council for Road Safety's ‘Road accidents’). Road traffic accidents arise from a multitude of intertwined factors, including road users’ behaviour, road infrastructure, vehicle conditions, and environmental elements. Understanding the factors that contribute to these road traffic accidents is crucial for developing effective preventive measures and ensuring safer road environments (N. Amarasingha's ‘Risk factors of crashes involving motorcycles in Sri Lanka’, S.K. Ahmed, M.G. Mohammed, S.O. Abdulqadir, R.G.A. El-Kader, N.A. El-Shall, D. Chandran, M.E.U. Rehman and K. Dhama's ‘Road traffic accidental injuries and deaths: A neglected global health issue’, A. Chand, S. Jayesh and A.B. Bhasi's ‘Road traffic accidents: An overview of data sources, analysis techniques and contributing factors’, D.S. Kodithuwakku and T.S.G. Peiris's ‘Factors contributing to the road traffic accidents in Sri Lanka’, and I. Ashraf, S. Hur, M. Shafiq and Y. Park's ‘Catastrophic factors involved in road accidents: Underlying causes and descriptive analysis’).
The significant advancements in the field of AI have shown promising results in solving complex problems in several identified domains. By harnessing the power of AI, we can potentially enhance our ability to identify and mitigate the factors leading to road traffic accidents, ultimately paving the way towards a safer future. While traditional methods have provided valuable insights into accident causation, the integration of AI technologies holds immense potential for revolutionising accident prevention strategies.
AI encompasses a range of techniques, including machine learning, computer vision, and natural language processing, which can be leveraged to process and analyse vast amounts of data efficiently. This enables the identification of patterns, trends, and risk factors associated with road traffic accidents that may not be readily discernible through conventional approaches (J. Siswanto, A.S.N. Syaban and H. Hariyani's ‘AI in road traffic accident prediction’).
Retrospective descriptive studies offer an invaluable approach to retrospectively examine past accidents, identify contributing factors, and establish correlations between variables. By leveraging data from postmortem reports, we can uncover critical information that shed light on the dynamics of road traffic accidents. Analysing these postmortem reports and death investigations-based information is essential for gaining comprehensive insights into accident causation patterns, focusing on the safety measures that can be adapted in futuristic vehicles that travel on the road. Though closed circuit television (CCTV) camera footage may provide a more detailed option to analyse traffic accidents, it may not be a practical technique in all situations.
Methodology
The 99 post-mortem reports belonging to the investigators and documents connected to death investigations due to road traffic accidents were selected from fatalities reported to the Colombo North Teaching Hospital (CNTH), Ragama, between 2001-2021. Data were gathered based on a pre-prepared questionnaire. Police sketch data was also used to further evaluate the road traffic accidents in selected cases.
Results
The majority of road traffic accidents reported to the CNTH and handled by the investigators involved men (83.8%), and 17.2% involved females. The majority of the victims belonged to the age group of 51-60 years (19.2%), followed by 61-70 years (18.2%), 41-50 years (16.2%), and 21-30 years (16.2%).
Pedestrians were the most vulnerable group (41.5%) among the road users, and motorcyclists (17.2%) were the second largest group facing road traffic accidents. The other vulnerable road users were passengers of vehicles, pillion riders and three-wheeler drivers.
Major roads with heavy traffic (72.7%) were the most frequent places for road traffic accidents, followed by junctions (16.2%), by-roads (6.1%), and bends (1%). Side impacts were the most common collision pattern (41.4%), followed by head-on collisions (28.3%) and rear-end collisions (18.2%). In the remaining 12.1% of the cases, details regarding the type of collision were not recorded.
The majority of accidents occurred between 6.01 p.m. and 12 midnight (38.4%). The other collision times were 6.01 a.m. to 12 noon, and 12.01 p.m. to 6 p.m. The accident rate was the lowest between 12.01 a.m. and 6 a.m. (9.1%).
The majority of the road traffic accidents (85.8%) had contributory human factors, based on the inference from the available history and other investigative documents submitted by the Police. It was difficult to infer positively about any documented contributory environmental, road-related, or vehicle-related factors with the available data. Following careful analysis of the available data by the investigators, cases were identified as being due to breaking road regulations (27.4%), poor judgement (23.6%), and overspeeding (7.5%). Poor lighting (5.1%) and road bumps (1%) are some road and environmental factors that may have contributed to road traffic accidents in selected situations. Other human factors contributing to road traffic accidents are intoxication, disabilities, and vision/hearing impairment.
Head injuries (41.4%) and multiple trauma (37.4%) were the most common causes of death, followed by chest injuries (8.1%) and neck injuries (1%). Other causes contributed to death in 12.1% of the cases. Blood alcohol was sent in 60.6% of the selected population, and 18.2% were found to be positive with the available reports (greater than 80 milligrams per decilitre [mg/dl]). Blood alcohol reports could not be traced in 14.1% of the victims.
Discussion
Middle-aged (51 to 60 years) male adults were the most frequent victims of road traffic accidents. According to N.W.M. Madhumali, H.M.T. Bandaranayaka, G.C. Hashari, H.K.B. Ayesha, R.M.K.P. Rathnayaka, I.U.M. Withana, F.M. Mowlana, P.L. Weerawardhana and J.M.K.B. Jayasekara's ‘Identification of risk factors for road traffic accidents using injured drivers: A cross sectional study conducted in Sri Lanka’, which was done at the Kandy National Hospital and the Kurunegala Teaching Hospital, the most common age group involved in road traffic accidents was those between 26-35 years, while 97% of the victims were male. T.M.F. Wazeema's ‘A matter of life and death: Road traffic accidents in Sri Lanka’ also found that the majority of the victims of road traffic accidents that occurred were male. This is presumably due to the fact of shouldering family responsibilities and thus frequenting motorways as pedestrians, motorcyclists, or occupants of other passenger vehicles.
The slight shift in the age group is based on the data available. In addition, it must also be noticed that although young adults are more prone to non-fatal crashes, the aged may have a higher death rate (K. Bhalla, K.V. Navaratne, S. Shahraz, D. Bartels, J. Abraham and S. Dharmaratne's ‘Estimating the incidence of road traffic fatalities and injuries in Sri Lanka using multiple data sources’).
According to global road safety data from the WHO in 2018, more than half of the worldwide road fatalities involve pedestrians, cyclists, and motorcyclists (‘Global status report on road safety’). W.N.S. Perera, W.A.S. Harshana, A.G.R.K. Bandara and P. Paranitharan's ‘A descriptive study on the pattern of regional injuries in fatal road traffic accidents in the CNTH’ also found that the high fatality rate among pedestrians (37.2%) was followed by bicycle riders or motorcyclists or pillion riders (44%). J. Edirisinghe's ‘Pedestrian-involved road traffic accidents – Case study in the Kandy Town’ also revealed that pedestrians involved in road traffic accidents are at a very high percentage. During the past decade, a significant elevation in the number of accidents involving motorcycles has been identified in Sri Lanka. According to the statistics for 2022 available with the aforementioned Ministry, the number of persons killed in road accidents mostly involved motorcyclists and pedestrians. The young age, the desire to drive motorcycles, and the easy availability of motorcycles at an affordable rate may have contributed to the increasing numbers.
Furthermore, unfavourable behaviour, mostly breaking road regulations by pedestrians and drivers (27.3%), may be the primary contributory human factor. Driving at excessive speed had been a contributory cause in 7.5% of the accidents. An Indian study proved that vehicles that had a speed range of 40-60 kilometres per hour (37.9%) were responsible for higher percentages of accidents. Further, 27.3% of the incidents were due to violations of road rules and regulations while walking and driving on the roads.
Driving under the influence of alcohol is another risk factor that promotes road crashes.
Alcohol intoxication was identified in 18.2% of the accidents, where the alcohol level was >80 mg/dl, based on the toxicology reports. Kodithuwakku et al. had observed that the consumption of alcohol is an important human factor contributing to road traffic accidents. The poor judgement of road users (23.6%) could be another main reason behind road traffic accidents. Poor judgement may have had a role in the causation of the accidents when opportunities were there to avoid a collision. The investigators were limited by a lack of visuals from CCTV footage, which may increase the accuracy of the interpretation; however, careful analysis of the data often identified poor judgement as a contributory human factor supported with sketch analysis.
In addition to that, the majority of the road traffic accidents occurred on roads with heavy traffic during the evening hours due to the large number of vehicles and pedestrians occupying the main roads compared to junctions and by-roads during this period, as well as the lower visibility of the surroundings. Madhumali et al.'s study also found that the majority of road traffic accidents had occurred after 6 p.m. and that 76.9% of road traffic accidents had occurred on carpeted main roads rather than on any other types of roads. Side impacts were the main method for collision identified, amounting to 41.4%, followed by head-on at 28.3% and rear-end at 18.2%. Faults in the braking system, defects in the engine, and a lack of modern safety features are the vital factors affecting road accidents. Also, environmental factors and road conditions are considered major contributory factors for road traffic accidents (per an Indian study). Weather conditions, including temperature, the wind speed, and humidity, represent the principal external factors that have an impact on road traffic crashes. Only a few incidents were identified due to environmental and road factors. This may be partly due to the lack of complete data available at the time of the death investigation.
Recognising the factors that contribute to road traffic accidents and taking preventive actions is essential to save valuable lives. The persisting problem of road crashes will not be solved merely by the strict implementation of road rules and severe punishments. Therefore, Thilakarathna et al. proposed potential remedies for reducing road traffic accidents using applications of AI. With the introduction of AI, it is rapidly transforming the vehicle industry and revolutionising road safety. Based on the most common vehicle used on Sri Lankan roads, Thilakarathna et al. proposed a motorbike that is developed using applications of AI.
Driver assistance systems (DAS) have become a sought-after technology with the advances in AI (F. Jiménez, J.E. Naranjo, J.J. Anaya, F. García, A. Ponz and J.M. Armingol's ‘Advanced driver assistance system for road environments to improve safety and efficiency’). Autonomous driving systems are one of the major applications in driver assistance systems (a proposed autonomous driving system has cameras that detect human emotions that are visible, biosensors that detect the alcohol level, blood pressure [BP], and the heart rate of the driver, light sensors that automatically detect the light intensity of the surrounding, and speed sensors that automatically detect the tyre speed and condition of the road surface and ignite the alert system). The incorporation of AI algorithms makes a vehicle more autonomous, resulting in self-navigation and perceiving and adapting to dynamic environments, resulting in a safer journey. In order to work without manpower, an autonomous system has been included with cameras to monitor the vehicle’s position within the lane and detect the driver’s mental condition while driving. It may be presumed that an emotionally unstable person, showing signs suggestive of fatigue, may be more prone to collisions, and thus, the driving can be switched to a driver assistance system. D.H. Lee and A.K. Anderson's ‘Reading what the mind thinks from how the eye sees’ found that a person’s emotions and feelings can be interpreted with more than 60% accuracy by analysing their eye expressions.
Based on that information, it is proposed that the autonomous motorbike system includes an eye-expression-sensitive camera on its helmet. Also, this system can be included with various sensors to automatically measure the speed of the vehicle and light intensity of the surroundings and provide real-time feedback to assist drivers in making safe decisions while driving. The sensors above a certain speed limit or very poor road surface may activate an alarm system to alert the driver, and similarly, when the environmental lighting is not sufficient, the intensity of the headlight and other supportive lighting systems can be activated. A.P. Selvam, S. Muthukumar, V. Kamakoti and S. Prasad's ‘A wearable biochemical sensor for monitoring the alcohol consumption lifestyle through ethyl glucuronide (EtG) detection in human sweat’ reiterates that the sweat alcohol levels correlate with the blood alcohol level.
Biosensors can be attached to detect the alcohol level, BP, and the heart rate of the motorcyclist. Based on the biosensor data, an autonomous system can recognise the physiological condition of the driver and alert the driver regarding their fitness to drive and may suggest automated driving to minimise collisions; basically, to follow road regulations more precisely and to minimise errors in judgement.
One of the most significant advances in road safety has been collision detection and prevention technology (in the proposed collision detection system, radar can detect the short- and long-range distances, and cameras capture the 360 degree angle view and road signs). Vehicles now use AI to detect potential collisions and implement preventive steps to avoid them. The motorbikes of the future can be facilitated with AI-powered collision avoidance systems with radar sensors to avoid objects in the path, including humans, by using an 80 gigahertz (GHz) wideband frequency modulated continuous wave (FMCW) radar sensor on a maximum measurement range of 5.69 yards with a movable radar target (L. Piotrowsky, T. Jaeschke, S. Kueppers, J. Siska and N. Pohl's ‘Enabling high accuracy distance measurements with FMCW radar sensors’). These systems can intervene autonomously by adjusting the motorcycle’s speed, steering, or braking to avoid or mitigate crashes.
The increasing urban population has become a major challenge in managing the traffic flow, resulting in road traffic accidents worldwide. To address these problems, cities are looking for smart traffic management systems that use AI algorithms to analyse real-time data from several sources, such as Global Positioning System (GPS) devices on vehicles (in the proposed traffic management system, the GPS system identifies the vehicle location and shares this information with other vehicles). R. Santiago and M. Ballesta-Garcia's ‘An overview of Lidar imaging systems for autonomous vehicles’ stated that GPS uses the signal from a network of satellites.
AI algorithms can identify traffic patterns, predict changes in conditions, and provide recommendations for optimal traffic settings. This helps reduce traffic congestion and improves the smooth movement of vehicles. It detects locations in real-time, designed to detect accidents rapidly, which can reduce emergency response times, remove reporting errors, and save more lives.
Conclusion
Road traffic accidents have become a major barrier in the development process of Sri Lanka, which highly impact all spheres of life. Road traffic accidents and deaths are a result of multiple factors, including human factors, poor environmental conditions, mechanical defects in vehicles, and demographic characteristics of the accidents. However, the majority of the contributory factors were human-related. This will prevent most fatalities and provide the complementary support needed to enhance road safety for a better future.