Wake up, have breakfast, brush your teeth, get ready for work, and head to the train station. As you open the door, someone from the house next door leaves at exactly the same time. A few hundred meters later, you find yourself walking alongside several other people, all moving in the same direction. By the time the train station comes into view, dozens of people are surrounding you.
You enter the station and people are everywhere. Rushing down the stairs, you encounter others climbing up, blocking your path. You bump into a few people, and finally manage to squeeze into the crowded train. The carriage is so packed that you cannot even reach your phone in your pocket. At the next station, even more passengers push their way inside. It feels impossible that there could still be room, yet at every stop more people enter, filling every gap.
Eventually, you arrive at the hub station. Streams of people move in every direction. Most of the time you simply follow the flow, but suddenly someone cuts across your path, forcing you to stop abruptly while the people behind bump into you.
Does this sound familiar? If so, you probably live in a large city, perhaps Tokyo like I do. Even if you live in a rural area and commute by car, or do not commute at all, you have likely attended a concert, festival, or sporting event where you could feel the warmth and pressure of the crowd around you.
Crowd management
Sometimes, being surrounded by people can feel exciting and comforting. People gather at stadiums, concerts, and festivals not only for the event itself, but also for the shared sense of belonging that comes from being among others who share the same passions. Yet crowds can also become dangerous.
During the 2022 Halloween celebrations in Itaewon, South Korea, more than 150 people lost their lives in a crowd crush. None of the attendees could have imagined such a tragedy. The location where the accident occurred was simply a narrow alleyway, similar to countless others found in cities around the world.
Can such tragedies be avoided? How can we understand, and perhaps even predict and modify, the collective motion of people?
To answer this question, let us return to the train station. When it was originally built, the surrounding area probably contained only a small number of homes, and its design was likely based on existing facilities of similar scale. As the population increased, the station became more crowded, requiring expansions and renovations. However, once buildings become deeply integrated into the urban environment, large-scale reconstruction is often difficult or impossible. In such cases, the only viable solution is to improve the efficiency and smoothness of crowd movement within the available space. This is what we call crowd management.
Effective crowd management requires the ability to quantify how people move. Incremental improvements are often guided by experience. Workers at train stations know their facilities extremely well, and changes usually occur gradually: a few more passengers each year, small adjustments over time, and continuous refinements to improve overall flow.
However, for entirely new facilities, such as airports costing billions of dollars and designed to accommodate millions of passengers, or for large-scale events, experience alone is not enough. Designers need quantitative criteria to determine what constitutes overcrowding and to establish safety margins.
Traditionally, crowd density plays this role. At densities below approximately two people per square meter, people can generally walk comfortably without significant interference. As a result, many public facilities are designed with this threshold in mind (or even lower limits, which is even better). Above four people per square meter, however, crowds begin to behave more like fluids: they fluctuate, compress, and can even become turbulent. Morning commuter trains in major cities sometimes reach densities of six people per square meter, a level that is potentially dangerous, though temporarily tolerated because train carriages are relatively small and staff closely monitor the situation.
Density is therefore an essential factor in crowd safety and facility design. Yet it also has an important limitation: it is fundamentally static. For example, when people move smoothly in the same direction along a corridor, relatively high densities may still pose limited risk because collisions are rare. By contrast, chaotic movement involving intersecting flows from multiple directions can create dangerous conditions even at moderate densities.
The congestion index
In my previous research, I developed methods to quantify the smoothness of crowd motion and identify problematic areas within public facilities that require improvement. The measure I developed, called the congestion index, is based on principles from fluid-dynamics and considers whether collective motion of people is stable (i.e. a so-called laminar flow) or unstable (a turbulent flow). Although mathematical details may be difficult to grasp, it is simplified by quantifying the smoothness of crowd motion on a scale from 0 to 1. A value of 0 represents perfectly coordinated, uncongested movement, while a value of 1 indicates highly chaotic and potentially dangerous conditions.
In another collaborative project, my colleagues and I developed sensors capable of measuring crowd density and walking speed in real time. The sensors were specifically designed for deployment during events and large-scale data collection campaigns, requiring only minimal installation and setup. Through years of research and collaboration with many colleagues and institutions, I eventually reached a point where I could both quantify the smoothness of crowd movement and collect real-time data to monitor crowd safety.
Once I became capable of identifying problematic areas in crowd flow, a new question naturally emerged:
How can we actively influence the motion of people?
This question is far more difficult to answer, and at present I still do not have a definitive solution, which is precisely why I continue to pursue research in this area.
Human movement is influenced both by external stimuli and by the physical environment itself. A brightly lit and visually appealing restaurant, for instance, may naturally attract more customers than a poorly lit restaurant next door. The same principle applies to public spaces. A clear sign can encourage people to move in a particular direction, even if it is not the shortest route to their destination.
To reinforce such guidance, staff members are often deployed to direct crowds. However, excessive intervention can negatively affect the atmosphere of a space. Consider Disneyland or other amusement parks: visitors do not want to feel constantly controlled or instructed. On the other hand, when safety becomes critical, subtle guidance is no longer sufficient, and roads or pathways may need to be physically closed to redirect people toward safer routes.
The concept of nudging
When only gentle influence is required, the concept of nudging may provide a solution. Lights, colors, sounds, and even smells can potentially guide people toward particular areas and improve the overall efficiency of movement. Although this may sound almost magical, its effectiveness is still not fully understood.
Can a simple light influence people to choose one direction over another? Does the effect change during daytime? Are brighter locations always more attractive? Can the same strategies influence groups of friends and families as effectively as individuals?
These are some of the questions my current research seeks to answer.
In particular, I am interested in understanding how imitation affects the success of nudges. Humans are deeply social creatures, strongly influenced by the behavior of others. If many people are lining up outside a restaurant, we instinctively assume it must be good. If the line becomes even longer, we may assume it is exceptional.
In this way, imitation may amplify the effects of nudges throughout a crowd. Yet many important questions remain unanswered: under what conditions does imitation emerge? When does it fail? How strongly do social relationships influence decision-making within crowds?
Finally, I also plan to work with animals to better understand these collective behaviors. Many social animals respond collectively to external threats or stimuli. Consider a flock of birds escaping from a predator: their movement is highly coordinated, with individuals simultaneously avoiding both the predator and collisions with one another.
By studying animal groups, I hope to gain deeper insights into how social systems, including human crowds, respond collectively to external pressures that require coordination among many individuals.
Ultimately, I hope this research will contribute not only to safer and more efficient crowd management, but also to a broader understanding of how societies collectively respond to challenges, from emergency evacuations and terror attacks to the spread of misinformation in our increasingly interconnected world.
