Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Current Hot Topics
  Autonomous Virtual and Artificial Characters: Realistic Inhabitants of Artificial Worlds  

Nadia Magnenat-Thalmann
University of Geneva, Switzerland and NTU, Singapore
Daniel Thalmann
EPFL, Switzerland and NTU, Singapore


1. Autonomous Virtual and Artificial Characters

Virtual characters, also known as virtual actors, have become very popular in the last 15 years, mainly through 3D movies and games. In movies, they now have very realistic physical and emotional characteristics, including their facial expressions, hair, clothes, and motions. In parallel, robotics has made considerable progress and new kinds of robots are available as social companions. Some of them have the appearance of humans or pets. Since they are physically present, we call them artificial characters rather than virtual characters. These social robots share a lot of research in common with virtual characters. In the rest of this essay, we will call them VACs (virtual and artificial characters).

VACs can play many roles, but we will distinguish two main ones: i-companions, which play a social role, and i-substitutes, which can stand in for a human in telepresence applications.

We envision that in the near future, i-companions, empowered with memory, emotions, and personalities, will live together with humans, entertaining and taking care of us. I-companions will ubiquitously shift form between physical robots and realistic 3D virtual humans. "I" stands for interactive, interpersonal, intimate, and individualized. I-companions will improve human quality of life by ensuring people assistance and safety anywhere, anytime. Deploying i-companions, at a fraction of the cost of actual humans, will alleviate the gross shortage of skilled human manpower, particularly in Western countries. Interactive and digital media have matured to the point that embodying these technologies in human characters like virtual humans and social robots is nearing fruition.

I-companions will play a major role in three application domains: education, security, and healthcare. First, i-companions can assist human educators of children. For example, while playing a serious game with a child, an i-companion might notice the child doing something dangerous, like approaching a hot stove. The i-companion could ask the child to stop, and, if necessary, call for human assistance. Second, i-companions will detect abnormal human behaviors in public spaces. In shops, sales i-companions will help customers find what they are looking for, and, at the same time, identify potential thieves. If a person exhibits unusual behavior, the i-companion will go and talk to her, to discern the person's true intentions. This is an important improvement over existing security technologies such as surveillance cameras, as i-companions can help prevent crime instead of just watching it happen. Finally, i-companions will monitor the sick or elderly, in homes and hospitals. A virtual nurse displayed on-screen in a hospital room will keep a patient company, while assimilating a staggering quantity and variety of data from cameras, microphones, and medical devices to ensure that the patient's condition remains stable. I-companions will sense and understand human behavior, assisting or watching over their human owners. An automatic understanding of social situations and human intentions is key to any of these scenarios.

Increasingly, people communicate remotely using new technical innovations, teleconferencing, and voice over IP, like Skype. These technologies allow people to attend meetings or gather no matter where they are. They can even share the same virtual space using 3D avatars, like in Second Life. More sophisticated systems are being developed with 3D capturing and rendering, leading to a true 3D telepresence experience, like in the BeingThere Centre. If the 3D avatar is distantly guided by the real participant, it is possible to replace this real participant with its autonomous virtual counterpart, as an i-substitute. The i-substitute is supposed to give a partial illusion that the real person is present. This implies that the i-substitute should look the same as the real human, speak with the same intonation, and be somehow aware of the real situation, the real participants, and the task being performed. The i-substitute should react at the right time based on its perception of the situation and the other (real) participants. It evaluates what each real participant is doing, and develops its perception from visual and audio input and recognition. The i-substitute reacts according to the input and its current knowledge. Its reactions encompass animation (body and facial gestures) and speech synthesis. The i-substitute could be an autonomous virtual human as well as a social robot, depending on the type of meeting and environment.

2. Properties of VACs

Autonomy is generally the quality or state of being self-governing. Rather than acting from a script, a VAC is aware of its changing virtual environment, making its own decisions in real time in a coherent and effective way. A VAC should appear to be spontaneous and unpredictable, making the audience believe that the character is really alive and has its own will.

To be autonomous, a VAC must be able to perceive its environment and decide what to do to reach an intended goal. The decisions are then transformed into motor control actions, which consist of animation for a virtual character and mechanical motion for a social robot, so that the behavior is believable. Therefore, a VAC's behavior consists of the following sequence: perception of the environment, action selection, and reaction.

The problem with designing VACs is determining how to decide on the appropriate actions at each point in time, to work toward the satisfaction of the current goal, which represents the VAC's most urgent need. At the same time, there is a need to pay attention to the demands and opportunities coming from the environment, without neglecting, in the long term, the satisfaction of other active needs.

There are four properties that determine how VACs make their decisions: perception and awareness, adaptation and intelligence, memory, and emotions.

Perception and Awareness. Perception of the elements in the environment is essential for VACs, as it gives them an awareness of what is changing. A VAC continuously modifies its environment, which, in turn, influences its perceptions. Therefore, sensorial information drastically influences VAC behavior. This means that we cannot build believable VACs without considering the way they perceive the real world and us. To realize believable perception, i-companions should have sensors (cameras, microphone) that play the role of the senses of real humans (eyes, ears). For social robots, these sensors can be directly built into the robot, but for virtual humans, we need to have these sensors as auxiliary devices, generally part of the display. What is essential for VACs is to recognize the people they meet and what these people do, which requires the complex processing of information.

Adaptation and Intelligence. Adaptation and intelligence define how the character is capable of reasoning about what it perceives, especially when unpredictable events happen. A VAC should constantly choose the best action so that it can survive in its environment and accomplish its goals. As the environment changes, the VAC should be able to react dynamically to new elements, so its beliefs and goals may evolve over time. A VAC determines its next action by reasoning about what it knows to be true at a specific time. Its knowledge is decomposed into its beliefs and internal states, goals, and plans, which specify a sequence of actions required to achieve a specific goal. When simulating large groups or communities of VACs, it is possible to use bottom-up solutions that use artificial life techniques, rather than top-down, plan-based approaches, such as those that are common in artificial intelligence. This allows new, unplanned behaviors to emerge.

Memory. It is necessary for a VAC to have a memory so that similar behaviors can be selected when predictable elements reappear. Memory plays an important role in the modeling of autonomy, as actions are often decided based on memories. But imagine a VAC in a room containing 100 different objects. Which objects can be considered memorized by the virtual character? It is tempting to decide that whenever an object is seen by the VAC, it should be stored in its memory. But if you consider humans, nobody is able to remember every single object in a room. Therefore, the memory of a realistic VAC should not be perfect either.

Emotions. The believability of a VAC is made possible by the emergence of emotions clearly expressed at the right moment. An emotion is an emotive reaction to a perception that induces a character to assume a physical response, facial expression, or gesture, or select a specific behavior. The apparent emotions of a VAC and the way it reacts are what give it the appearance of a living being with needs and desires. Without them, a VAC would just look like an automaton. Apart from making them appear more realistic, VACs' visible emotions can provide designers with a direct way of influencing the user's emotional state. It should be noted that emotions can be generated for recent social robots as their faces have deformable skin and can display subtle expressions and even lip synchronization.

3. The Impact of Research on Virtual and Artificial Characters

The four properties above are very important in creating believable VACs. Modeling these properties accurately in real time requires research efforts from various branches of computer science. We emphasize a few of them below.

Behavior Planning. Behavior planning involves the selection of appropriate actions for the VAC to execute. These decisions should reflect the individual characteristics of the VAC, including its intelligence, its motivations, and its social behavior. Besides being individual, action selection architectures for VACs should be both reactive and proactive to be efficient in real time. The transitions between reactions and planning should be rapid and continuous in order to elicit coherent and appropriate behaviors in changed or unexpected situations. The design of a behavior planner satisfying these criteria is a research challenge. If virtual characters and social robots share most of these properties, it is obvious that the motion of virtual characters will be smoother, but social robots will be able to realize concrete tasks like carrying real objects.

Sensor Design. VACs have to be aware of events and characteristics of the real world. It takes real devices, such as cameras, microphones, and haptic devices, to capture this information and bring it to the virtual characters. The information they provide must be integrated with that of virtual sensors.

Modeling Emotions. To allow VACs to respond emotionally to a situation, they could be equipped with a computational model of emotional behavior. Emotion-related behavior, such as facial expressions and posture, can be coupled with this computational model, which can be used to influence their actions. The development of a good computational model is a challenge.

Modeling Attention and Gaze. When we walk into a city, we look at other people, objects, or even at nothing in particular. An important aspect that can greatly enhance the realism of crowd and group animation is for characters to be aware of their environment and of the other characters in it. When adding attention behaviors to crowds, we are confronted with two issues: detecting the points of interest the characters are looking at, and editing the characters' motions for use in modeling the gaze behavior.

Modeling Memory-based Emotions. Some researchers mathematically model emotions, behavior, mood, and personality for virtual characters. These models can be used to create an emotionally responsive VAC. However, such models lack the critical component of memory--a memory of not just events but also of past emotional interaction. A memory-based emotion model is needed to take into account the memory of past interactions in order to build long-term relationships between the virtual character and users.

4. Ethical Concerns

There are ethical concerns regarding the use of VACs. One such concern involves decisions made by real people that are based on the advice of autonomous characters. Autonomy means that the VAC makes decisions based on his or her understanding of the environment and of the rules for the surrounding world. In a simulated world, how can we be sure that this information corresponds to reality? As with any computer program, VACs are not immune to bugs or tampering. Their advice on critical matters should always be validated.

Another ethical concern is the fact that an autonomous virtual human may be indiscernible from a real existing person. For example, a terrorist group could create a TV spot showing democratic leaders promoting nondemocratic values. To avoid misleading and manipulating the public, we will need to use technology, such as watermarking, to reliably indicate to the viewer that the human is virtual.

Some autonomous characters promote violence, terrorism, abuse, or crime, in the context of games or other interactive situations. If their behavior is realistic, they are likely to exert a strong negative influence, even when they are known to be virtual. New laws and regulations will have to be developed in this area.

When avatars and VACs are together in the same virtual community, it becomes very difficult for a member of the community to know when he or she is interacting with an avatar or a VAC. In the near future--when VACs can replace avatars when the user is away--the problem could become even trickier.

5. Scalability and Mobility

With the advent of wearable devices, advanced PDAs, and smartphones, VACs can be with people all the time to guide and help them. Moreover, with light see-through glasses, it becomes easy to add virtual characters to a real scene. Applications of this technology could be to show people how to use electronic devices or how to find their way through a particular area without the use of a map. If a social robot is used, it can show the way or even help to carry objects.

6. Possibilities and Challenges

VACs will be essential in many applications of the future. They will be our playmates, teachers, therapists, and pets. Because of their logic and memory, autonomous characters will bring skills and abilities that complement those of humans, rather than replace them. The next generation--children and young adults--is very open to virtual communities and e-learning, and is expected to have no problems interacting with virtual humans.

How far are we from such a situation? Current VACs are becoming more realistic in terms of their appearance and animation, and they are able to perceive the virtual world and the people living in that world. They may act based on their perception in an autonomous manner. New social robots have an increasingly realistic human appearance and can have smooth facial expressions and gestures. However, their intelligence is constrained and limited.

In the near future, we may expect to have VACs that are able to learn or understand a few situations, due to the development of new methods of artificial intelligence and behavior. However, a great deal of research effort is still needed to reach the point at which VACs can behave autonomously and interact naturally, like real humans. True simulation of the full complexity of human behavior will take more time.


Created: Feb 14 2006
Last updated: Feb 12 2014


Web Pages

The Institute for Media Innovation: housed at Nanyang Technological University (Singapore); the Core Group works on both virtual humans and social robots.

BeingThere Centre: International Research Centre for Telepresence: a joint center between Nanyang Technological University (Singapore), the University of North Carolina (USA), and the Swiss Federal Institute of Technology (Switzerland).

The Center for Human Modeling and Simulation: investigates computer graphics modeling and animation techniques for embodied agents, virtual humans, and their applications.

MIRALab: a pluridisciplinary lab at the University of Geneva that is working on virtual human simulation and virtual worlds.

CMU Social Robots Project: a project with the goal of making the communication and interaction with robots easy and enjoyable.


An immersive multi-agent system for interactive applications Wang, Y.; Dubey, R.; Magnenat-Thalmann, N.; Thalmann, D. The Visual Computer 29, 5 (2013), 323-332.

Anthropomorphism of artificial agents: a comparative survey of expressive design and motion of virtual characters and social robots Dalibard, S.; Magnenat-Thalmann, N.; Thalmann, D. In Proc. of the Autonomous Social Robots and Virtual Humans Workshop (Singapore, May 9, 2012).

Making them remember -- emotional virtual characters with memory Kasap, Z.; Benmoussa, M.; Chaudhuri, P.; Magnenat-Thalmann, N. IEEE Computer Graphics and Applications 29, 2 (2009), 20-29.

Simulating gaze attention behaviors for crowds Grillon, H.; Thalmann, D. Computer Animation and Virtual Worlds 20, 3-4 (2009), 111-119.

Intelligent virtual humans with autonomy and personality: state-of-the-art Kasap, Z.; Magnenat-Thalmann, N. Intelligent Decision Technologies 1, 1-2 (2007), 3-15.

An integrated perception for autonomous virtual agents: active and predictive perception Conde, T.; Thalmann, D. Computer Animation and Virtual Worlds 17, 3-4, (2006), 457-468.

Generic personality and emotion simulation for conversational agents Egges, A.; Kshirsagar, S.; Magnenat-Thalmann, N. Computer Animation and Virtual Worlds 15, 1 (2004), 1-13.

The Thing Growing: autonomous characters in virtual reality interactive fiction Anstey, J.; Pape, D.; Sandin, D. In Proc. of the IEEE Virtual Reality Conference 2000 (VR '00) (New Brunswick, NJ, Mar. 18-22, 2000), 71-78.


Crowd simulation (2nd ed.) Thalmann, D., Musse, S. R., 2012

Handbook of virtual humans Magnenat-Thalmann, N., Thalmann, D. (Eds.), 2004


Let's keep in touch online: a Facebook aware virtual human interface Liu G., Choudhary S., Zhang J., Magnenat-Thalmann N. The Visual Computer 29 (9): 871-881, 2013

Modelling and controlling of behavior for autonomous mobile robots Skubch H., Springer Vieweg, 2013

Building long-term relationships with virtual and robotic characters: the role of remembering Kasap Z., Magnenat-Thalmann N. The Visual Computer 28 (1): 87-97, 2012

Introduction to autonomous mobile robots (2nd ed.) Siegwart R., Nourbakhsh I., Scaramuzza D., MIT Press, 2011

Stepping into virtual reality Gutierrez M., Vexo F., Thalmann, D., Springer, 2008

Virtual storytelling: using virtual reality technologies for storytelling Subsol G. (Ed.), Springer, 2006

Fast multi-level adaptation for interactive autonomous characters Dinerstein J., Egbert P. ACM Transactions on Graphics 24 (2): 262-288, 2005


Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright © 2000-2017 ThinkLoud, Inc.
Terms of Use
| Privacy Policy