The Age of Robotics
Technology has evolved to extend our senses, as with telescopes, different types of imaging and means of communication at a distance. It also extended our human capabilities with control of power, as in flight and other means of transportation. Robotic technology can add to or even replace our very presence in the world, and the impact of robotic systems in our lives, especially when coupled with artificial intelligence, will be pervasive, possibly beyond what we can even imagine.
There are really 3 distinct areas of robotics, each with its own impact: industrial robotics (often identified with “automation”), telerobotics, and autonomous robotics. We will consider these, together with related technologies.
Industrial robotics is the most developed area. It is applied to repetitive tasks, often with high levels of precision, and has largely taken the place of assembly line workers. That’s how cars are built, as are most manufactured objects produced in large quantities. These systems may or may not be “programmable” for different tasks. When each task is “fixed” the systems are more appropriately referred to as “automation”. Robotics is usually associated with the idea of “programmability”, i.e. the possibility of instructing the same machine for a range of different tasks.
The first example of such a system was the Unimate that welded parts on a General Motors car assembly line. The machine could be programmed for different tasks from a drum memory (computers were not around yet). This was demonstrated on The Tonight Show, hosted by Johnny Carson, where the machine drove a golf ball into a cup and performed other simple feats.
Industrial robotics is now at the heart of most manufacturing and the machines are programmed and controlled by computers.
Telerobotics provides the ability to manipulate objects at a distance under direct human control. It is made possible by control systems that can simulate human actions at different levels of resolution, ranging from the Canadian arm manipulating large objects in the vicinity of the International Space Station, to the Da Vinci Surgical Robot developed in 1999 and capable of performing sutures and delicate tissue manipulation. Robonaut, developed for space applications, is another example of a system that can manipulate common tools with hands that simulate human hands. Teleoperated robots are particularly useful in environments where human skills are needed but which are too dangerous for people, such as disaster areas, nuclear reactors, handling of explosives, etc. The aim of the technology is to provide the user with visual and sensory feedback as well as control ability that ideally can substitute for human presence.
Autonomous robotics is the area that most excites the imagination, in that it envisions machines that can act of their own accord just like living creatures. This type of robotics is made possible by the development of Artificial Intelligence, and the tasks possible for robots are limited by the progress of Artificial Intelligence and the particular needs that are being addressed. For instance, the intelligence requirements for a robotic vacuum cleaner like the Roomba are like those of a simple insect, and are handled by current AI with relative ease. Creating a “humanoid” robot that could interact rationally with people, as seen in science fiction, is still a distant goal. We’ll return to the humanoid idea below.
Most autonomous robotics is task-specific and requires different types of intelligence, ranging from pattern and object recognition to the ability to navigate through obstacles and manipulate objects as needed. The Mars robotic rovers provide a good example in that they need to be able to recognize and avoid rocks and crevices, while moving on their own, as signals to Earth and back can take around 40 minutes, so any human involvement as in telerobotics would be impossible.
Probably the greatest short-term impact of autonomous robotics will be caused by the advent of autonomous vehicles (AV). also known as self-driving cars. These are in fact “robots”, in that they sense the environment and make decisions in how they interact with it and with other objects, be they other AVs, people, or obstacles. The same technology, which is already beginning to be introduced, will also become part of aviation. Autonomous robotic systems will prove to be much safer than humans, as they will only improve in time, while we are already acting at the highest level of our error-prone capabilities.
Biologically Inspired and Modular Robotics
We have already followed the evolution of organisms over billions of years. We were mainly tracking the evolution of our own species and its ability to develop science and technology, but the “biological technologies” created and adopted by every single organism over the same period of time are often still unparalleled and remain a source of inspiration and, whenever possible, useful imitation. The flight of insects and birds, and legged locomotion are good examples. While at NASA I collaborated with a German scientist (Frank Krishner) on a project for developing a dog-sized 8-legged robot we called “Scorpion” that would be able to go into rough terrains where wheeled robots cannot. We envisioned it as an adjunct to the current Mars rovers. For a similar purpose I led a “Snakebot” project: a robot built of small modules connected together and hinged in a way that allowed for snake-like locomotion. There are numerous examples of such robots of all sizes built by several research institutions, and inspired by different biological properties. The Snakebot was also an example of “modular robotics” and that is robots built of several identical modules, or with connecting interfaces that allow for different configurations.
The idea of building an “artificial human being” has always captured human imagination. One of the first models of a moving human body, with gears and pulleys was actually designed, again, by the extraordinary genius Leonardo da Vinci. Of course, there were no sources of power, nor computer intelligence to make it work, until now. A good example is Asimo built by Honda. The problem is that artificial intelligence is able to perform very difficult tasks only in separate and well-defined areas. There is still nothing that can approach the ability, complexity and integrated abilities of the human brain, so humanoid robots, and, similarly “pet” robots like Aibo, developed by Sony, remain entertaining curiosities. What is probably a more important development is the possibility of interfacing robotic prosthesis, like limbs and internal organs, with our body and brain. We will address the implications of these technologies later in the course.
Share with friends