As one of the most outstanding young experts in the field of robotics in China, he has been deeply involved in robot research and teaching for more than ten years, and personally participated in and witnessed the whole process of robot technology from germination, cultivation, development to wide application in China.
He is Li Xinde, a staunch explorer and a pure scientist in the field of robotics.
He said that scientific research is not only intensive cultivation on the basis of predecessors, but also a brave exploration of unknown fields. For the future robot world, we need to constantly break through and make bold innovations.
"There is a nursing robot that can be used in nursing service application scenarios such as elderly care."
"There is an intelligent welding robot that can perform automatic welding, which greatly reduces the labor intensity and saves labor."
"There is an amphibious robot whose design concept is to combine multi-legged and flapping wings to achieve land and air use."
…………
Professor Li Xinde from the School of Automation of Southeast University is usually reticent. However, whenever it comes to artificial intelligence or robots, he is like a "chatterbox", as if he is a different person and has all kinds of whimsy. In his laboratory, all kinds of robots and systems mentioned above are being designed intensively.
Not long ago, as an important member, he participated in the project "Key Technologies and Applications of Complex Positioning and Navigation Control of Mobile Medical Service Robots" and won the first prize of scientific and technological progress of China Automation Society. This project is aimed at the major needs of national development, especially the major anti-epidemic needs. Focusing on the three key technical problems of mobile service robot, such as complex environment perception, autonomous planning and positioning, and intelligent navigation control, the key technology research of mobile service robot is carried out by combining Industry-University-Research. The research in these fields is Li Xinde's "specialty". During the epidemic prevention and control in COVID-19 in 2020, the autonomous medical service robot developed by using this core technology will perform disinfection tasks in hospitals, nursing homes, rehabilitation centers, communities and other places, greatly reducing the workload of medical staff, reducing the risk of cross-infection, ensuring the normal life of the people during the epidemic prevention and control, and achieving extremely high social and economic benefits.
In fact, in the past ten years, Li Xinde has undertaken more than 20 national and provincial scientific research projects, authorized 17 national invention patents and 5 software copyrights, and achieved considerable results and awards. However, compared with various honors brought by scientific research, Li Xinde is more concerned about the difficulty of his own challenges. In recent years, he has devoted himself to solving cutting-edge, world-class professional problems, forgetting to eat and sleep, working day and night, and never letting him give up even if he encounters bottlenecks.
Li Xinde comes from a small village in Yantai. At that time, if a rural child wanted to change his life track, there was only one way, reading. When he was a sophomore, he took part in a liberal arts class. After he was promoted to the third year of high school, he felt that he was not interested in liberal arts and transferred to the science class. This is an extremely risky choice, because the learning content of liberal arts and science is very different. In order to successfully cross the "single-plank bridge" of college entrance examination, Li Xinde made use of all available time to catch up on science courses, and at the same time sprinted to review with the teacher's pace. Finally, through his own efforts, he was successfully admitted to the major of chemical equipment and machinery in shenyang institute of chemical technology. It can be seen that Li Xinde is a person who is unwilling to step by step and has a strong spirit of challenge. This quality laid a good foundation for him to constantly challenge himself in academic research and achieve one success after another.
During college, in order to save living expenses, Li Xinde only ate two meals a day. Even the lack of material conditions does not affect his determination to learn knowledge. Due to outstanding academic achievements and performance, he was rated as a school-level outstanding graduate when he graduated. After graduation, in order to relieve the financial pressure of her parents, sensible Li Xinde did not choose to pursue further studies, but worked as a technician in a profitable enterprise in Jinan for three years.
In the eyes of relatives and friends in my hometown, this job is an enviable and stable "iron rice bowl", but Li Xinde is still not satisfied. He has higher requirements for himself and a greater vision for the future. He decided to apply for a master's degree. At that time, the work unit did not support the in-service staff to take the postgraduate entrance examination. He only goes to work during the day, gets on his broken bike after work at night, and flies all the way to Jinan University for self-study at night without eating dinner. At that time, the roads in the suburbs of Jinan were potholes, and there was no street lamp for a section of the road. On one occasion, in order to save review time, Li Xinde rode his bike too fast and accidentally fell into the manhole cover with a loose manhole cover. He was injured in many places and a large piece of skin was worn off his face.
The hardships of life did not stop Li Xinde's footsteps, but sharpened his mind and carved a rough jade to shine. Hard work pays off, and finally he was admitted to the Master of Mechanical Design and Theory of Shandong University. During his years of further study in Shandong University, he was favored by Professor Song, a famous person in the field of mechanical vibration, and was selected as his disciple with his solid theoretical foundation and indomitable learning attitude. Under the influence of Professor Song's teaching and subtle influence, my knowledge and insight have increased a lot, and I have also achieved fruitful scientific research results in my study, and won the highest scholarship "Principal Scholarship" of Shandong University.
From the whimsy of youth to the hobbies of adulthood, artificial intelligence and robots have always been Li Xinde's persistent dream. After entering the 2 1 century, China's artificial intelligence has gone from the foggy initial stage to the breakthrough of imprisonment. At that time, Li Xinde, with his foresight, keenly grasped that artificial intelligence would be the ultimate technical field in the next decade or two. Therefore, he made up his mind to apply for the Department of Control Science and Engineering of Huazhong University of Science and Technology and successfully became a doctoral student of Professor Huang. This year, he was 28 years old. In Li Xinde's bones, he inherited the honesty and tenacity of Shandong people, and he didn't want to talk about the hard work, because in his view, it was an opportunity of life to participate in such a grand torrent of rapid development of artificial intelligence and contribute his meager strength to the country.
His tutor, Professor Huang, once served as a member of the former "863" national expert group on intelligent robots, and has been struggling in the front line of artificial intelligence research and often participated in academic exchanges at home and abroad. Under the teacher's demonstration, Li Xinde got many opportunities to exercise, which made him more interested in artificial intelligence and automation, and also made more profound and detailed scientific research work. Of course, there are always difficulties. The first is the knowledge gap brought by cross-majors. Automation and artificial intelligence majors are quite different from mechanical majors, and there are many courses to make up. In addition, Professor Huang is a famous strict teacher in the industry. If he can't meet his academic requirements, he will never give any accommodation.
After careful consideration, Li Xinde thinks that the only way he can persist is to redouble his efforts. He chose the most "stupid" way, such as insisting on getting to the laboratory as early as possible every morning and leaving at the latest in the evening, so that the research time in the laboratory can reach at least 12 hours; In his spare time, he spent almost all his time reading materials in the library and searched almost all the materials about artificial intelligence. It is relying on such a two-part reference document and a two-part professional book that the drops of water converge into the sea and gradually accumulate their own unique knowledge in this field. He discusses automation and robotics with his tutor and fellow students every day, asking for advice with an open mind, and thinking about the progress of his discipline all the time.
Li Xinde is always at the forefront of artificial intelligence, and opportunities always favor those who are prepared. By chance, Li Xinde read the newly published The Paradoxical Theory (DSmT), was deeply fascinated by it, and tried to study it deeply. He met Jean Dezert, a French scientist, and Professor Florentine Smarandache of the United States, the founder of the Paradox Theory (DSmT) and an internationally renowned expert in information fusion. After long-term exchanges and discussions, Li Xinde forged a profound friendship with them. Among them, Jean Dezette was invited to visit Southeast University twice, and Smaranda of Florence awarded him the certificate of international scientific contribution in the name of his unit because of Li Xinde's special contribution to specious theory.
Li Xinde's stupid way of "fighting a dark war against a hard wall" made him unconsciously become one of the best students in the tutor team. Generally speaking, it takes 5 6 years to complete his doctoral studies. He graduated only four years ago, and when he graduated, he also won the highest scholarship of Huazhong University of Science and Technology-"Rockwell Scholarship". Because of his outstanding performance in professional research, many well-known universities in China have thrown olive branches to him. In the face of the choice of "squandering flowers to make your eyes charming", Li Xinde has his own opinions, and he finally chose Southeast University, a century-old institution.
"Southeast University is a century-old university with a reputation as an' academic holy land'. It is one of the earliest departments to offer automation specialty in China, and it is also one of the earliest universities to carry out robot research in China. It has a solid foundation and strong strength for the construction of artificial intelligence. I think this platform is very suitable for me! " Facts have proved that Li Xinde made the right choice again.
For a long time, robot has been an important research field of artificial intelligence. Since the invention of the first robot in 1950s, scientists have been working hard to study all kinds of robots that can imitate human and biological movements for more than 60 years, and the robot technology has been continuously improved. In the past, many imaginations about robots in science fiction works have gradually become a reality.
However, we have to admit that robots are not out of the scope of automatic machines. People and robots are still defined as the relationship between use and being used, substitution and being replaced, and robots appear more as a functional replacement tool. The close coordination and cooperation with people in the same space are rarely considered in the design process of robots, which makes the basic problems such as intrinsic safety, human-computer cooperative cognition and behavioral mutual assistance not well solved, and the flexibility and intelligent decision-making ability are still insufficient.
Integration, cooperation, * * enjoyment and * * governance are the remarkable characteristics of the future intelligent society. With the rapid development of artificial intelligence technology, intelligent robots gradually enter human life and strive to become an indispensable member. In communities, hospitals, nursing homes, factories ... people expect robots to have multiple roles and perform interdisciplinary tasks in different scenarios. However, if robots want to achieve intelligent tasks, communicate seamlessly with human beings and blend harmoniously, they still need a revolutionary breakthrough in the ability to perceive, understand and interact with the environment.
Li Xinde concluded that at present, intelligent robots face three main problems in location perception and understanding, namely, lack of integration and utilization of multi-source information of high-complexity locations, lack of effective clues to identify locations and lack of semantic interpretability in the perception process, which leads to low recognition rate of current methods and is difficult to integrate into intelligent tasks. In order to overcome the above shortcomings and effectively expand and extend human capabilities, the future development direction of robots-* * * fusion robots came into being, but to achieve the goal of * * * fusion, multi-level innovation of robot technology is needed.
The so-called "* * * fusion robot" refers to a robot that can naturally interact with the working environment, people and other robots, adapt to the complex dynamic environment independently and cooperate.
Li Xinde emphasized that "* * * integration" contains three meanings: the first layer is the natural interaction between robots and the environment. Traditional robots work in specific areas with simple targets and single objects, but it is difficult to have such a pure environment in actual production and life. Let the robot walk out of the workshop, into the snow in Shan Ye, into the busy streets, and even let the robot and people work together in a dynamic environment, which is the first goal of the robot.
The second is the mutual assistance and complementarity between robots. Two people can lift heavier objects, and a team of three F 1 racing service personnel can change a tire in two seconds. Complex medical operations require the cooperation of many doctors and nurses. How can robots achieve a breakthrough in single ability through multi-person cooperation? This is the second goal of the development of * * * integrated robots, to realize the cooperation and complementarity between robots, and to improve the overall function and performance of robots through cooperation.
The third is the cooperative work between robots and people. To cooperate with others is not only to listen to their instructions, but also to have their own "judgments" and "ideas". People are complex and dynamic, and cooperation with people is a higher requirement for robots. At this time, the robot is no longer passively waiting for human assistance or blindly relying on human programming, but should be able to actively perceive and solve various problems. This requires robots to be able to perceive their own environment, so as to judge what tasks they should perform. On the other hand, in the era of man-machine integration, in order to make robots and humans communicate seamlessly, the expression of target semantics and place semantics in natural language interaction is very common. When robot/unmanned system is oriented to interaction and work based on these concepts, it is difficult to understand and perceive these semantic concepts. "For example, if you go to the ward to look after patients and go to the kitchen to boil water, the robot should not only be able to perform specific actions or tasks such as looking after and boiling water, but also be able to analyze what kind of places the ward and kitchen refer to. Li Xinde said, therefore, the intelligent robot's perception of the place and its further understanding of the place play an auxiliary role in the decomposition of intelligent tasks and the good interaction between man and machine.
The reason why human beings can adapt to various environments is inseparable from their keen perception of different environments. * * * Fusion robots also need this kind of diverse perception. "It is worth noting that diversified sensing is not as simple as adding a few sensors, and there are many tasks to be done such as signal transmission, operation and identification. In many cases, to identify the external environment and the intentions of others, it requires the fusion of multiple modal signals, which requires the robot to perceive, understand and even learn in real time. " Li Xinde particularly emphasized.
Supported by the National Natural Science Foundation's major research project, Li Xinde led the team to work hard and put forward a unified deep neural network with multi-modal fusion. From the perspective of "prototype" clustering of position semantics, semantic map correlation reasoning and multimodal fusion interactive correlation reasoning, the position perception and understanding algorithm of robots and autonomous motion control of robots are deeply studied. Aiming at how to enhance the semantic interpretability of place perception process, from the perspective of natural language description, this paper puts forward the definition of place concept under semantic clues, and studies the corresponding recognition methods. Aiming at the semantic understanding and recognition of places under complex environmental information, a method based on the combination of images and their natural language representation is studied and proposed from the perspective of simulating human cognitive process of places, and the related image description generation technology and multimodal fusion technology are studied. From the perspective of place understanding framework, a place recognition and understanding framework based on knowledge map reasoning is studied and proposed, and the construction method and reasoning model of place domain knowledge map are given. Based on this, the Guangdong Pearl River Talent Innovation Team Project "Research and Development of Key Technologies and Industrial Application of Multi-task Scenarios and Cross-domain Unmanned Systems" was successfully declared and approved.
Li Xinde's in-depth study of this project has laid a solid foundation for further carrying out site-related tasks such as human-computer interaction, service strategy, task-level planning, autonomous operation, etc., and for intelligent robots to smoothly enter human family and daily life, and to communicate seamlessly and live in harmony with human beings.
At present, all the powerful countries in the world are playing games around the core technology of artificial intelligence. Whoever runs fast can seize the commanding heights first. From this perspective alone, the research of artificial intelligence is of great significance, which is related to the strategic deployment of building an innovative country and a world science and technology power.
In recent years, China has included robots and intelligent manufacturing in the priority areas of national scientific and technological innovation, and made great efforts to plan and make solid progress.
Li Xinde is eager to accept difficult challenges, and at the same time he can calmly analyze the situation. In his view, in the field of artificial intelligence and robotics, compared with foreign counterparts, China's disciplines really started late, but with the economic development, the country's capital investment and attention have been significantly improved, and there is a trend of catching up, and the gap with the international community is narrowing.
Over the years, Li Xinde has made all his efforts and painstaking efforts to root all his knowledge in the soil of the motherland and to make artificial intelligence and robot technology flourish in China. Li Xinde often thinks: "Science and technology workers need to have a sense of social responsibility. They should do some research and development and experiments with the feelings of benefiting the country and the people, make first-class achievements, solve practical problems of some major national projects and improve the living standards of ordinary people. It is the dream of science and technology workers! "
As an explorer and practitioner of robotics, Li Xinde said so and did the same. His wisdom and efforts are playing a role in many fields related to the national economy and people's livelihood.
-Facing intelligent manufacturing and modern logistics, such as disorderly sorting, intelligent welding, dynamic assembly and other robot application fields, robots require dynamic and high-precision three-dimensional measurement of the work site and the shape and posture of the workpiece. This poses a new challenge to the structured light sensor, which requires its high accuracy, small size, high speed, strong environmental adaptability and strong anti-interference ability. In the past, the equipment system of intelligent welding was completely monopolized by foreign countries, and the high cost of several hundred thousand yuan seriously hindered the development of the industry. However, with the support of the key scientific and technological achievements transformation project in Jiangsu Province-"Welding Robot Complete Equipment", Li Xinde team independently developed a set of structured light vision sensor and seam tracking system, which broke the monopoly of foreign high-priced tracking equipment market and created huge social and economic benefits. Supported by the key project of Jiangsu R&D plan-"Key Technology and Application of High-precision Structured Light Dynamic Measurement System Based on MEMS Chip", Li Xinde team initiated and opened a new era of 3D stereoscopic vision dynamic perception.
-In the field of health services, according to WHO, there are 400 million mental patients in the world. Mental illness has become an important disease in 2 1 century, and depression is now the fifth major cause of people's loss of mobility. However, it is gratifying that the daily actions and behaviors of patients with early depression can well reflect their symptoms, making it possible for machine vision to monitor and analyze their behaviors. Therefore, Li Xinde team uses non-contact visual perception and deep learning technology to intelligently analyze people's behaviors, emotions and mental states, so as to estimate people's mental states. This research is supported by national key projects. Aiming at the intelligent monitoring scene, this research achievement obtains the non-contact information of the human body in the natural state (interactive/non-interactive) by using the camera arranged in the intelligent environment or on the robot, and combines the prior knowledge and the understanding of the scene/place, and then analyzes the mood and depression state by fusing the multimodal information such as expression, micro-expression, eye movement and gait, and even by fusing the gaze and interactive behavior of the characters during the interaction, which is a precedent for the application of depression technology. The research results also lay a solid foundation for the service robot to provide high-quality service through online state analysis of the service object.
-In the field of national defense security, Li Xinde has carried out in-depth cooperation with many scientific research institutes and undertaken a number of national defense pre-research projects. It has made rich research achievements in small target detection, identification and tracking, individual abnormal behavior detection, group abnormal behavior detection, violent behavior detection, multi-machine cooperation, group intelligence, air-ground cross-domain cooperative operations, etc., and has broad application prospects in national border defense, street fighting, reconnaissance assault, security and other application fields.
Artificial intelligence is not only accelerating development, but also upgrading. It is a strategic technology that leads the future. Countries all over the world have realized that the future will be the key stage of competition among countries, and regard the development of artificial intelligence as a major strategy to enhance national competitiveness and safeguard national security.
The huge population and complete industrial structure provide China with the potential to create massive data and a broad market. But China still needs to do a lot of basic work, such as a more open data environment and well-trained scientific talents. In particular, continuous efforts in talent will be the decisive factor.
"Whoever has first-class innovative talents and first-class scientists will have an advantage in scientific and technological innovation." Li Xinde naturally understands the importance of team strength and talent echelon construction. In fact, he has been rooted in this fertile soil for more than ten years since he was invited by Southeast University in 2007. He is not only engaged in scientific research, but also shoulders the heavy responsibility of cultivating talents.
In Li Xinde's mind, his relationship with students is not only teachers and students, but also friends, and he is a teammate who will lead the world in the field of automation research in China in the future. Li Xinde showed his unique affinity and frankness towards team members. He prefers to hear students call him "Lao Li" rather than "Miss Li" or "Professor Li".
Li Xinde often said that every student is unique. "Never underestimate your students, they are all capable." This is a sentence that Li Xinde often talks about. He will teach students in accordance with their personality and characteristics, and always tap their potential.
"The most important thing is not your learning ability, but your ability to analyze and solve problems and your ability to innovate." Li Xinde repeatedly warned the team members, "Your future competitors are not people, but machines. You tell it to learn, it is faster than you; You are more diligent than it, and it is more diligent than you. So what do you PK with it? Must be the ability to innovate. You must have a forward-looking vision, not just staring at the land in front of you. "
In order to broaden students' horizons and increase their knowledge, Li Xinde will provide them with many opportunities to participate in various meetings, exchanges and lectures, so that they can exercise and encourage them to go out bravely and breathe more "fresh air" if they have the opportunity to study and communicate at home and abroad.
"Going abroad is a good thing to broaden your horizons, and it is also a very important part of the growth of young people, which will make their thinking more active." At that time, Li Xinde visited Georgia Institute of Technology and National University of Singapore as a visiting scholar and postdoctoral fellow, from which he got to know Professor Donald Akin, the father of American robots, the winner of the Best Invention Award of Time magazine, a famous robot expert and Professor Ge Shuzhi, an academician of Singapore Academy of Engineering.
"They are all people with great wisdom, especially Mr. Ge. Our chat is not limited to academic aspects, but also involves philosophy of life. Communicating with people who stand tall and look far will give me a lot of inspiration. " Li Xinde sighed.
From single-handedly, poor, to increasingly large and powerful Mazhuang, Li Xinde led this fast-growing team into the scientific research highway. The management mode of Li Xinde team is very open. Although he is the "top leader" of the team, he never talks about everything. Instead, he set up a management Committee, which was divided into several groups according to different projects and different directions, and all major and minor issues were decided by the Committee. Li Xinde's knowledge and personality have been unanimously recognized and respected by the team members. Under his leadership, a robot automation iron army with strong cohesion and combat effectiveness was born. In 20 19, the postgraduate team directed by Li Xinde went out for the first time and won the special prize of the Target Detection and Recognition Challenge of China Aviation Society.
"Last year's epidemic had a great impact on us. We plan to complete six projects. Due to the suspension of classes in the first half of the year, we basically discussed various problems in the implementation of the project online. " Li Xinde's memories of last year are still thrilling. "But at the critical moment, the team demonstrated its combat effectiveness, worked hard to overcome various difficulties, and finally completed the task on time and with high quality!" Li Xinde's words revealed his pride in the team. Under his leadership, such a team with enthusiasm for exploration, innovative spirit and international vision has made solid progress step by step on the road of scientific research and achieved good results.
However, Li Xinde is not satisfied with its achievements. "On the one hand, in scientific research, if we want to solve a specific problem, we must encourage the cross-integration of teams, and we need to be brave enough to do things we have never done, climb the peak of science and lead the trend of scientific and technological development; On the other hand, in engineering application, we should integrate theory with practice and pay attention to the transformation and application of Industry-University-Research's achievements. At this time, we must break through the boundaries of our own abilities, learn, and communicate and cooperate with people who have mastered new knowledge. " In his view, only in this way can researchers themselves, even a team, become stronger and stronger.
There are different opinions about the future development of the ever-changing artificial intelligence and robotics. "In fact, there is no way to accurately predict." Li Xinde changed the subject. "Scientific research is not writing a script, but a brave exploration of the unknown world. Maybe it will succeed, maybe it will fail. Isn't the fun of researchers in the process of exploration? "
Li Xinde, a professor and doctoral supervisor of Southeast University, is the vice chairman of the Intelligent Robot Professional Committee of China Artificial Intelligence Society, and the vice chairman of the Intelligent Products and Industry Working Committee of China Artificial Intelligence Society. Long-term commitment to machine vision, machine perception, intelligent robots, human-computer interaction, unmanned systems and so on. He has been selected into Jiangsu Blue Engineering Talent Training Plan, Jiangsu Liufeng Talent Training Plan, Southeast University National Jieqing Training Plan, Jiangsu Key R&D Plan, Chief Scientist of Key Projects, Guangdong Pearl River Talent Innovation Team, Jiangsu University Excellent Innovation Team, etc.