By Ganesh Chaudhari, Technical Manager, Sumo India Studios
The dawn of the 21st century has undoubtedly brought seismic changes in people’s personal and professional lives. Ever since the global pandemic, digitalisation gained centre stage, with the gaming paradigm shifting from PCs and console physical games to digital and online.
Given the rise of new-age technologies such as artificial intelligence (AI), machine learning, and virtual reality (VR), trends in the online video games industry are increasing at an unprecedented rate – driving sharp expansion in the sector. According to KPMG, India is predicted to become one of the leading markets in the industry, with a $3.9 billion market by 2025.
Furthermore, to establish a distinct position in today’s competitive market, gaming companies – and developers are constantly aiming to create realistic and lifelike virtual worlds that capture players’ imaginations. In this context, the simulation of realistic crowds within these game environments has emerged as a critical component to reaching this goal.
Though crowd simulation in the realm of games is not a new concept, and with technological advancements in recent years, the term has gained significant prominence in the industry. So, let us delve into the intriguing concept of crowd simulation in games by exploring the gaming world’s techniques, obstacles, and effects.
Crowd Simulation: A Fascinating Aspect in the Gaming Industry
Crowd simulation is an incredibly fascinating aspect of the gaming industry. It involves the creation, and simulation of a substantial number of ‘agents’ or crowds in a believable manner throughout the virtual world. The simulation is a cumulation of multiple layers of techniques coming together.
Though the concept has come a long way since its inception, today’s games feature some impressive crowd simulations that can make players feel like they are part of a living, breathing world.
One of the most exciting things about crowd simulation is its potential for creating truly immersive gaming experiences. Whether you are playing a game set in a bustling city or exploring a vast open world, the ability to interact with and observe large groups of characters can add an incredible level of depth to the gameplay.
Creating a convincing crowd simulation requires multiple iterations of tuning to the underlying systems, a task that can go back and forth between the designers and programmers. The goal here is not to create a realistic crowd but to create a believable one.
A good simulation might not always behave realistically (e.g. a commuter choosing a busier road instead of a sparse one) but it should never behave erratically (e.g. the commuter exiting his vehicle and starting to walk because they saw a red light at a junction).
The art of crowd simulation has advanced over time, and contemporary games today aim to produce complex and lifelike crowd behaviours.
Techniques involved in crowd simulation
We have seen crowd simulation which is as simple as every member of the world randomly walking around the streets, to the complexity of every person in a city having a concrete alibi, vocation, and a purpose for being where they are in games past and present. Regardless of the complexity, a crowd is simulated in multiple layers, each adding to the complexity and credibility of the output/simulation.
Here, we will discuss a few of the most common layers of tech in crowd simulations.
Navigation and Movement: We can consider crowd simulation as a subset of game AI. And just like any intelligent agent in the game who can traverse the world they are in; the crowd too must be able to perform the most fundamental operations of pathing and locomotion.
The pathing aspect here deals with being able to chart an optimal route/course for a crowd member to follow facilitating free movement along the world. This involves the application of different pathing strategies and pathing algorithms.
The pathing aspect often relies heavily on the game’s navigation system. Making sure that each virtual character can move around the game world efficiently is not enough, we must also make sure that they do it in a believable manner by avoiding dynamic obstacles seen on the route, which often will be other people.
This is where obstacle avoidance comes into play. It can be thought of as fine adjustments that are being made to the charted course to make sure the crowd navigates tight or crowded spaces naturally.
Furthermore, to create a realistic and engaging gaming world, the crowd must be animated in a realistic and natural way. By leveraging technology, characters can have realistic motions, gestures, and reactions by using animation techniques like motion capture, procedural animation, or a combination of the two.
Awareness and Emotional Touch: Virtual characters need to have a sense of sensibility and awareness to make the crowd simulation seem more realistic. They need to be able to respond to their environment, identify and avoid challenges, and react to the activities of other characters.
To simulate spatial awareness – the ability to understand the ‘lay of the land’ – tools such as environment query systems and spatial partitioning are often employed. Perceptual awareness can also be modelled so that a crowd member only reacts to local events that fall within their perception range.
Furthermore, crowd simulation can incorporate character emotions and social interactions to boost realism and awareness further. Virtual characters can display more human-like behaviours and responses by incorporating models of emotions, personalities, and social relationships.
Crowd Behaviour Modelling: Establishing a set of guidelines or behaviours that govern how characters interact with one another and their surroundings is known as crowd behaviour modelling. These guidelines include avoiding collisions, adhering to predetermined pathways, responding to changes in the environment, and adopting sensible social behaviours.
This last aspect is not at all the least of them. In fact, it is quite important how social behaviours are modelled within the crowd. For example, it defines the density of the crowd at a particular zone in the game, at a particular time of day.
It also includes modelling the interactions between one or more crowd members. These interactions can be modelled by keeping them exclusively scripted to allow them to be fully emergent. Normally, you would want to find a sweet spot between those extremes.
Crowd Variation and Performance Optimisation: Simulating a considerable population can be expensive computationally. Various optimisation approaches can be used to guarantee real-time performance to cope with the effect.
To maximise the performance of crowd simulation systems various optimisation tools such as simulation and rendering LODs (level-of-detail), instance meshing, time slicing, load balancing and parallel processing are often used. A crowd should be diverse in terms of looks, actions, and demographics to be considered lifelike.
Game designers can use procedural generation techniques to add variety to character features, including appearance, attire, and personality traits. This can further enhance crowd simulation and aids in avoiding repetitive patterns.
Crowd Simulation: Gaming Experience Redefined
In a bid to gain more traction, gaming sectors have implemented and adopted various trends, such as crowd stimulation to provide immersive and interactive experiences in the virtual world. In fact, crowd simulation is not limited to just human characters; it can also involve simulating crowds of animals, vehicles, or other entities in the game world.
This has not only enhanced the gameplay for gamers but has also helped developers and gaming studios to create ultra-realistic games to deliver immersive and interactive experiences. By leveraging the power of artificial intelligence, animation, and crowd simulation, a new world of lifelike gaming has emerged, changing the landscape of the modern gaming world.
