Perspective| From Automatic Control to Intelligent Robots

  • Share:

Robotics is a broad and interdisciplinary subject that involves automatic control, embedded computing, sensor fusion, navigation, map building, planning and artificial intelligence. Its utmost goal is to develop intelligent robots that can think and function as our humans do. This is a very challenging and ambitious task since intelligent robots must cope with various noises, uncertainties, and dynamical changes in the real world. Like human beings, robots should be able to sense their environments, reason and make decisions, and respond to tasks and unexpected events autonomously. This has inspired scientists to provide robots with a fast connection from sensing to action autonomously during the last few decades. 

 

With the development of classical control theory in the 1950s, robots entered a stage of practical applications. The world's first robot used in industrial production could perform trajectory control with a high accuracy. In the 1970s, industrial robots were deployed rapidly as modern control technology and computer technology were successfully developed. These industrial robots have multi-joints, multi-CPU control systems, and multiple sensors such as tactile and force sensors. In this period, robots only had the ability of memory and storage, and conducted routine operations in structured manufacturing environments according to the pre-programed procedures. They cannot operate in dynamic environments with unexpected changes.

 

In the 1980s and 1990s, the fast advancement of artificial intelligence and sensing technology enables robots to operate in unstructured and unknown environments. By using various sensors, including visual sensors and non-visual sensors, such as force, touch and proximity sensors, the robots could obtain real-time information about their working environments and the objects to be manipulated. Behaviour-based approach was invented for robots to operate autonomously in dynamic and unknown environments. In this period, robots had some simple intelligence to conduct the required tasks and could handle dynamical changes in their surroundings. Therefore, the application of robots was extended from the manufacturing industry to non-manufacturing industries, such as mining robots, construction robots, hydropower system maintenance robots, even defence and military robots.

 

Since the beginning of the 21st century, the significant development of intelligent control and artificial intelligence technologies has made robots autonomously operate in human-centred environments and provide services to humans in needs. This kind of robots has good environmental perception ability and is also able to interact with humans via voice and gestures. They can do simple reasoning, decision making, planning and learning, working independently according to operation requirements and environmental conditions. Also, they can be our friends or companions to support us, as well as our co-workers to complete some tasks collaboratively. Their application range is continuously expanding to diversified fields, e.g., medical robots, security robots, and housekeeping service robots.

 

We must also realise that the vigorous development of the robot industry is inseparable from the research progress and technical support of automatic control systems. In this regard, the current cutting-edge control techniques for future robots include the flexible robot control, bio-myoelectric control, sensitive tactile robot control, conversational interactive robot control, emotion based control, ideas-mind based robot control, autonomous driving, virtual reality based control, cloud robot control and so on. More specifically, these control techniques can be elaborated further below:

 

The flexible robot control technique can accurately control the appearance and shape characteristics and/or movement status of materials (e.g., liquid metals) used in robots. The soft robots being able to control their appearance and shape arbitrarily within a certain range have wide application prospects in the field of pipeline fault inspection, medical diagnosis, and disaster rescues. 

 

The bio-myoelectric control technique adopts the electromyographic signals on the surface of human upper limbs to control robotic arms. The bio-signal control based robots are largely deployed in the medical application and rehabilitation. 

 

The sensitive tactile control technique enables robots to sense the shape, texture and hardness of an object for manipulation. The robots with this technique are able to do a series of complex tasks such as medical treatment and exploration.

 

The conversational and interactive control technique makes robots understand user's questions and give accurate answers. It can also actively guide the completion of the conversation when the information is incomplete. 

 

The emotion based control technique enables robots to effectively recognize human emotions and psychological activities so that they can provide services in time. The robots with mental activity can be utilised in robot-assisted medical rehabilitation and criminal investigation. 

 

The ideas-mind based control technique employs brain-computer interfaces for collecting and recognizing the human brain signals to control a robot. It has been applied to the rehabilitation of the disabled people. 

 

The autonomous driving technique provides humans with a new transportation tool, which can optimize the road traffic autonomously. Most importantly, it can greatly reduce the number of road accidents and save more lives.

 

The virtual reality based control technique can effectively support the remote operation of the robot, such as in the fields of maintenance detection, entertainment experience, on-site rescue, military reconnaissance and so on. 

 

The cloud control technique enables robots to access data and knowledge available in the cloud and to receive commands via communication networks for their remote operation. The cloud robots can be accessed by their users 24 hours a day with minimum maintenance. They provide users with a new type of services that are infinitely scalable and used on demand.

 

Finally, we should emphasise that most of the autonomous robots that have currently been built do not function well in the real world, and their capabilities are far from our dream and expectations. This is mainly attributed to the limitations of current automatic control methods, the incompatibility of available computer technology, and our incomplete understanding of the perception, recognition and reasoning processes in humans. However, our dream will come true as our research efforts continue into the next century. Although it is very difficult and often unwise to predict the future, we believe that we are gradually progressing towards the robots that are intelligent and well-functioned in the real world as faster computers and new control technologies become available. It is no doubt that we will progress much faster as more and more talented young students and researchers join the team year by year.

 

Origin:

https://www.springer.com/gp/campaign/control-robotics/liu-guoping-hu-huosheng-blog

 

 

◇◇◇◇◇◇

For more up-to-date information:

1) WeChat: IJAC

2) Twitter:IJAC_Journal

3) Facebook:International Journal of Automation and Computing

4) Linkedin: Int.J. of Automation and Computing

5) Sina Weibo:IJAC-国际自动化与计算杂志

  • Share: