Research background
After the first wave (agricultural revolution) and the second wave (Industrial revolution), the third wave - the information revolution - is unfolding. The information revolution mainly includes two aspects: information production and information processing. In the period of information revolution, the rapid differentiation of science and technology disciplines, the increasingly detailed division of labor in various industries of social production, and the emergence of a large number of various sciences and technologies have formed a situation of "information explosion", and the development level of information processing has gradually become the bottleneck of the development of information revolution. Once the society achieves important results in information processing, it will often trigger and promote profound changes in the economic, political and cultural fields of the whole human society. Therefore, the development level of computer technology can be regarded as the central clue of the wave of information revolution, and the development level of artificial intelligence (including artificial emotion) can be regarded as the core content of computer development. Emotional robots are the result of a growing trend.
At present, artificial intelligence is showing rapid growth and comprehensive expansion. On the one hand, artificial intelligence continues to develop in the direction of deeper intelligence: mathematical operations, logical reasoning, expert systems, pattern recognition, deep learning, etc. On the other hand, it continues to expand to all areas of society: smart TV, smart phone, smart home, smart transportation, smart shopping, smart city, smart pension and so on. The next technological breakthrough of artificial intelligence is bound to be artificial emotion, and only when artificial emotion is realized in the true sense, artificial intelligence will have a broader development space and form a more powerful driving force for social productivity.
Emotional robot is the use of artificial methods and technology to give computers or robots with human-like emotions, so that it has the ability to express, identify and understand joy, sorrow and anger, imitate, extend and expand human emotions, which is the dream of many scientists. Compared with the high development of artificial intelligence technology, artificial emotional technology has made little progress. Emotion has always been an unbridled gap between human brain and computer. For a long time, emotional robots were the stuff of science fiction and rarely included in the research of scientists.
At present, countries around the world have invested a lot of manpower, material resources and financial resources in "artificial emotion", and have made a lot of progress in facial expression processing, emotional speech processing, posture expression processing, natural human-machine interface and so on. So far, artificial emotion has achieved only two things: (1) emotion expression system, and (2) emotion recognition system. However, the human emotion system includes at least five basic parts: the emotion expression system, the emotion recognition system, the emotion operation system, the will operation system, and the perceptive emotion interaction system. Obviously, all the research on artificial emotion to date has completely omitted the latter three systems.
In fact, emotion is a special way of thinking of human beings, which is completely different from the intelligent way of thinking, but the research and development of emotional robots must be based on scientific emotional theory, which is realistic. Without a new scientific emotional theory as guidance, it is impossible to develop emotional robots in a real sense. This new emotion theory must break through the limitations of psychology and social science, and become an independent, cross-scientific theory across natural science and social science, its fundamental purpose is: emotion digitization. This new emotion theory is "mathematical emotion", which takes "unified value theory" as the theoretical premise and uses mathematical logic method to analyze the emotion phenomenon and emotion law science. To sum up, "unified theory of value" and "mathematical emotion" mainly complete the theoretical framework of emotional robots through the following steps.
Research content
First, the research steps of emotional robots
1. Establish a new theory of value: the unified theory of value
This theory of value must be based on the axioms of natural science, unify the value theories of all disciplines (such as philosophy, economics, political economy, sociology, value engineering, etc.), and be able to fully reveal the internal logical relationship between various forms of value (such as money, real estate, status, health, appearance, talent, etc.). A mathematical model for calculating all kinds of values is established. Taking "dissipative structure theory" of physics as the theoretical premise, it studies life phenomenon and life law from the perspective of physics, and value phenomenon and value law from the perspective of energy, and realizes the unification, objectification and mathematics of value theory, thus promoting the natural scientific process of the whole social science, and also lays a theoretical foundation for the establishment of "mathematical emotion science".
2. Develop a new theory of emotion: mathematical emotion
The study of emotional phenomena and emotional laws from the perspective of axiology has realized the unification, objectification and mathematics of emotional theories, thus promoting the natural scientific process of emotional theories. It goes far beyond the scope of psychology, analyzes emotional phenomena and emotional laws from the perspective of value, makes mathematical definitions of values and emotions, and deduces the "three laws of emotional intensity". It also realizes the mathematical operation of different emotions and lays a theoretical foundation for the creation of "artificial emotions".
3. Build a new artificial emotion theory: emotional robots
The logical block diagram of eight important systems is drawn: emotion expression system, emotion recognition system, emotion operation system, will operation system, perceptual and emotional interaction system, memory system, attention system, self-awareness system, etc., and a preliminary technical scheme is proposed for the development of artificial emotion.
4. Realize the industrial revolution of artificial emotion: Artificial emotion +
After the preliminary technical scheme of artificial emotion is completed, an emotion chip can be developed according to its basic ideas, and it can be implanted into an intelligent robot, and a true emotional robot can be produced. At the same time, the emotion chip can also be widely used in various industries, thereby triggering the whole society's "artificial emotion +" industrial revolution, which is a broader and profound industrial revolution after the "Internet +".
In short, without "unified axiology" there is no "mathematical emotion", without "mathematical emotion" there is no "emotional robot", without "emotional robot" there is no "artificial emotion +" industrial revolution.
Second, the technical route of emotional robots
The process of interaction between human and the outside world can be divided into five relatively independent stages: feeling stage, cognitive stage (or analysis stage), evaluation stage (or emotion stage), planning stage (or will stage), and behavior stage. Each addresses five questions of a different nature: What is there? What is it? What's the use? What to do? Kill it! They are completed by five main executive systems (sensory system, cognitive system, evaluation system, volitional system and behavioral driving system) and three auxiliary executive systems (memory system, attention system and self-awareness system). The logical sequence of information processing is: external information, sensory information analysis, cognitive information evaluation, behavioral information evaluation, will information implementation.
It can be seen that the technical line of emotional robot research can be divided into four aspects: the technical line of the research object, the technical line of solving the problem, the technical line of executing the system, and the technical line of information processing.
1. Technical route of the research object
The technical route of the object of study: the material attribute of the thing a the relation attribute of the thing a the value attribute of the thing a the value attribute of the behavior a the implementation mode of the behavior.
2. Technical route to solve the problem
The technical route to solving problems: What is it? What is a? What's the use of a? a What should I do? a Kill it.
3. Execute the technical route of the system
(1) The technical route of the main executive system, namely, sensory system a, cognitive system a, evaluation system a, volitional system a, behavioral system;
(2) The technical route of the auxiliary executive system, namely, the self-awareness system a memory system a attention system.
4. Technical route of information processing
Technical route of information processing: external information a analysis of sensory information a evaluation of cognitive information a evaluation of behavioral information a implementation of volitional information.
The technical route of emotional robot research.
Equipment recommendation
Emotional robot
1. Have visual and voice interaction ability, can observe, listen, express, operate, and realize interaction ability such as greeting and chatting with people.
2. Highly anthropomorphic: with anthropomorphic shape, the expression robot has both facial actions, and the body action robot has both body actions and facial actions
3. Emotional cognition: The eyes are equipped with CCD cameras, and the expression changes of the communicators are observed after determining the brightness of the room
4. Will elaboration: After receiving the stimulus signal, determine the intensity and direction of emotion, select the mode of emotion expression, receive and process the feedback signal generated by emotion expression
Expressive robot
Limb action robot
reference
S. Schaal, S. Kotosaka, D. Sternad, "Nonlinear Dynamical Systems as Movement Primitives", Proc. Humanoids 2000 IEEE-RAS Int'l Conf. Humanoid Robots, 2000.
J.G. Hale, F.E. Pollick, "'Sticky Hands' Interaction with an Anthropomorphic Robot", Proc. 2000 Workshop Interactive Robotics and Entertainment, pp. 117-124, 2000.
G. Tevatia, S. Schaal, "Inverse Kinematics for Humanoid Robots", Proc. IEEE Int'l Conf. Robotics and Automation, pp. 294-299, 2000.
T. Martinetz, K. Schulten, "Topology Representing Networks", Neural Networks, vol. 7, no. 3, pp. 507-522, 1994.
M. Kawato, "Internal Models for Motor Control and Trajectory Planning", Current Opinion in Neurobiology, vol. 9, pp. 718-727, 1999.
S. Vijayakumar, S. Schaal, "Fast and Efficient Incremental Learning for High-Dimensional Movement Systems", Proc. IEEE Int'l Conf. Robotics and Automation, pp. 1894-1899, 2000.
A. Ude, C.G. Atkeson, M. Riley, "Planning of Joint Trajectories for Humanoid Robots Using B-spline Wavelets", Proc. IEEE Int'l Conf. Robotics and Automation, pp. 2223-2228, 2000.
M. Riley, A. Ude, C.G. Atkeson, "Methods for Motion Generation and Interaction with a Humanoid Robot: Case Studies of Dancing and Catching", Proc. 2000 Workshop Interactive Robotics and Entertainment, pp. 35-42, 2000.
T. Shibata, S. Schaal, "Fast Learning of Biomimetic Oculomotor Control with Nonparametric Regression Networks", Proc. IEEE Int'l Conf. Robotics and Automation, pp. 3847-3854, 2000.
S. Schaal, C.G. Atkeson, S. Vijayakumar, "Real-Time Robot Learning with Locally Weighted Statistical Learning", Proc. IEEE Int'l Conf. Robotics and Automation, pp. 288-293, 2000.
C.G. Atkeson, S. Schaal, "How Can a Robot Learn from Watching a Human?" , Proc. 14th Int'l Conf. Machine Learning, pp. 12-20, 1997.