Gender and Age-Based Violations in Motorcycle Behaviours: A Case Study in Marrakesh
This paper presents a deep learning-based system utilizing YOLOv11 and CNN models to detect specific motorcycle traffic violations in Marrakesh while analyzing their correlation with rider age and gender to enhance urban road safety and law enforcement.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the bustling streets of cities across the developing world, a quiet revolution in transportation is reshaping how people move. Motorcycles have surged in popularity, offering a nimble and affordable alternative to cars, especially in places where traffic congestion is a daily reality. In Marrakesh, Morocco, these two-wheeled vehicles now make up more than sixty percent of all road users. While this shift brings efficiency, it also introduces a complex web of safety challenges. The unique way riders weave through tight spaces and navigate crowded intersections creates hazards not just for themselves, but for pedestrians and drivers of larger vehicles. To understand and improve safety, researchers are turning to a field of science that allows machines to see and interpret the world much like humans do. By using cameras and advanced software, scientists can now watch traffic flows, identify specific behaviors, and even recognize the people behind the handlebars. This approach moves beyond simple counting; it seeks to understand the human element of traffic, looking for patterns in how different groups of people interact with the rules of the road.
A team of researchers from Cadi Ayyad University in Marrakesh has taken this concept a step further by building a system designed to spot dangerous habits among motorcycle riders. Their work focuses on four specific infractions that frequently lead to accidents: riding without a helmet, using a mobile phone while driving, running red lights, and zigzagging erratically between other vehicles. To study these behaviors, the team set up a high-resolution camera at a busy intersection known for its heavy motorcycle traffic. Over the course of several days, they recorded more than seven hours of video footage, capturing the daily rhythm of the city during morning, afternoon, and evening hours. This raw visual data became the foundation for their investigation, allowing them to observe real-world driving habits as they happened.
The core of their method relies on a sophisticated type of artificial intelligence known as deep learning. The researchers first used a powerful detection tool to scan the video frames and locate every motorcycle and rider. Once a vehicle was found, the system zoomed in to identify the driver's face. From there, the software analyzed specific facial features to estimate the rider's age and determine their gender. This demographic information was then linked directly to the actions the rider was performing. The system was trained to recognize when a helmet was missing, when a hand was holding a phone near the face, when a vehicle crossed a stop line against a red signal, or when a motorcycle moved in a jagged, unpredictable path between cars. By connecting who the rider was with what they were doing, the researchers could move beyond general statistics to see if certain types of people were more likely to break specific rules.
The results of this analysis revealed clear and distinct patterns in rider behavior. The data showed that age is a significant factor in how likely a person is to commit a violation. Riders between the ages of eighteen and twenty-five were the most prone to using mobile phones while driving, with this group showing the highest rates of distraction across both genders. In contrast, the twenty-six to thirty-five age group appeared more frequently in the data for failing to wear helmets. When it came to gender, the study found that male riders generally committed violations at higher rates than female riders across all categories. This was particularly evident in the dangerous practice of zigzagging between vehicles, where young men aged eighteen to twenty-five accounted for nearly a third of all such incidents. Female riders, especially those in older age groups, were observed breaking the rules far less often.
To make sense of these findings, the researchers built a decision model that mapped out how age, past behavior, and gender interact to influence a rider's choices. The analysis suggested that age is the most powerful predictor of risky behavior. Younger riders were more likely to take risks, while older riders tended to be more compliant. The model also indicated that a rider's history matters; those who had broken rules in the past were more likely to do so again, suggesting that unsafe driving is often a consistent habit rather than a one-time mistake. While gender played a role, its influence was not as strong as age and was often dependent on the specific context of the violation. For instance, while young men were the most frequent offenders overall, there were specific situations where the difference between male and female behavior was less pronounced.
The study concludes that these intelligent systems offer a promising way to enhance traffic safety without relying solely on human observation. By automatically detecting violations and linking them to the demographics of the riders, authorities can move toward more targeted enforcement and education strategies. Instead of applying a blanket approach to all drivers, safety campaigns could be tailored to address the specific risks faced by young men or the particular distractions affecting the eighteen to twenty-five age group. The researchers note that while their system worked well with the data they collected, future improvements could include factors like weather conditions and traffic density to make predictions even more accurate. Ultimately, this work demonstrates that by understanding the human patterns behind traffic violations, cities can develop smarter, more effective ways to protect everyone on the road.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.