Cognitive Robotics

 What is cognitive robotics? This seems like an easy question to answer: cognitive robotics is at the intersection of robotics and cognition. Or: Cognitive Robotics is at the intersection of robotics and cognitive science. However, what does that intersection look like? What, exactly, is the relationship here? Below we will three different ways of looking at this intersection.

1. Probably most intuitive, cognitive robotics is about doing robotics that deals with cognitive phenomena such as perception, attention, anticipation, planning, memory, learning, and reasoning. Now, some people believe that robotics already deals with those phenomena, and are therefore left wondering how cognitive robotics would be any different from robotics, period. However, despite what you see in the movies, most existing robots don't learn, have no memory to speak of, and don't reason. In fact, at this point most existing robots are used in industry (think assembly lines), and most of them don't even have any perceptual abilities at all; they are programmed to do one thing, and one thing only. This kind of robotics we might call Industrial Robotics, and it can be characterized with the 3 D's of robotics: robots that do dull, dangerous, or dirty work, that no human would or can do ... which is exactly why Industrial Robotics is important! However, it is not what we see as Cognitive Robotics. In Cognitive Robotics, we are interested in the kind of robots that are, well ... more cognitive. Robots with the kind of intelligence that humans have. Robots that reason, remember, learn, and that can communicate with humans and with each other. Robots that can be characterized by the 3 C's: Clever, Creative, and Charismatic.

2. Creating such cognitive robots is obviously not an easy task. The field of Artificial Intelligence should clearly be a field we could use here, and in the cognitive robotics courses we teach, and in the Cognitive Robotics research we perform, in the Cognitive Robotics lab we certainly make use of AI techniques. However, another strategy might be to try and have a robot perform tasks the way humans do. That is, we could take our best theories and models from Cognitive Science, and try to apply and implement them in our robots. This could actually be a somewhat different way of looking at Cognitive Robotics, i.e. as the application of cognitive science to robotics. As an example, consider a robot that needs to catch a ball. You might think that a robot would solve this task as follows: take a snapshot (or couple of snapshots), determine the location, direction, and speed of the ball, compute when the ball is going to be where, and compute how to move the arms, legs, and all the other joints of the robot to be at that spot at the right time. Cognitive science research, however, has shown that when people catch a ball, they probably use quite a different strategy, the basic idea of which is that if the ball moves to the left in your field of vision, then you move the left, and if it moves to the right, you move to the right. Of course, this strategy requires one to keep looking at the ball (or at least frequently look), but that is exactly what a typical human does. But the point of the example is this: Cognitive Robotics could be seen as the doing of robotics, informed by Cognitive Science.

3. Of course, as we actually implement Cognitive science theories or models in a robot, we may find that the robot doesn't perform as we thought it would, meaning that maybe our theory isn't as good as we thought it was. In this sense, we can also turn things upside down: instead of cognitive science informing or helping robotics, we can regard the doing of robotics as informing cognitive science. For example, if we have two competing explanations or models for how humans perform certain cognitive tasks, then we could possibly implement each of those models in a robot, and see which robot more closely mimicks human performance. This way, robots can be used as a testbed for cognitive science theories, which is a third way to think about Cognitive Robotics. Finally, though, why should a robot be constrained by doing things the way humans do things? If a robot can do things better or more effective than a human by using a different kind of strategy, isn't that ok? And yes, of course, for most practical purposes that should be indeed be perfectly ok. But notice that even in that case, robots can be used to inform cognitive science. How would that work? Well, the trick is to regard cognitive science as the science of all of cognition, not just human cognition. Indeed, if you think about it, human cognition probably only takes up a very small spot in the whole space of cognition, and a true cognitive science will therefore have to consider kinds of cognition quite unlike human cognition. Well, robots could be a great way to explore those other kinds of cognition. As such, cognitive robotics could be considered a kind of Experimental Cognitive Science.

So what is cognitive robotics? Above we have seen several different ways to look at it: as the creation (engineering) of robots with cognitive abilities, as the creation of such robots using the knowledge of cognitive science, and finally as using robots to inform the field of cognitive science. In our courses, research, and lab, we look at cognitive robotics in all these ways.Cognitive Robotics or Cognitive Technology is a subfield of robotics concerned with endowing a robot with intelligent behavior by providing it with a processing architecture that will allow it to learn and reason about how to behave in response to complex goals in a complex world. Cognitive robotics may be considered the engineering branch of embodied cognitive science and embodied embedded cognition, consisting of Robotic Process Automation, Artificial Intelligence, Machine Learning, Deep Learning, Optical Character Recognition, Image Processing, Process Mining, Analytics, Software Development and System Integration.

Core issues

While traditional cognitive modeling approaches have assumed symbolic coding schemes as a means for depicting the world, translating the world into these kinds of symbolic representations has proven to be problematic if not untenable. Perception and action and the notion of symbolic representation are therefore core issues to be addressed in cognitive robotics.

Starting point

Cognitive robotics views human or animal cognition as a starting point for the development of robotic information processing, as opposed to more traditional Artificial Intelligence techniques. Target robotic cognitive capabilities include perception processing, attention allocation, anticipation, planning, complex motor coordination, reasoning about other agents and perhaps even about their own mental states. Robotic cognition embodies the behavior of intelligent agents in the physical world (or a virtual world, in the case of simulated cognitive robotics). Ultimately the robot must be able to act in the real world.

Learning techniques

Motor Babble

Main article: Motor babbling

A preliminary robot learning technique called motor babbling involves correlating pseudo-random complex motor movements by the robot with resulting visual and/or auditory feedback such that the robot may begin to expect a pattern of sensory feedback given a pattern of motor output. Desired sensory feedback may then be used to inform a motor control signal. This is thought to be analogous to how a baby learns to reach for objects or learns to produce speech sounds. For simpler robot systems, where for instance inverse kinematics may feasibly be used to transform anticipated feedback (desired motor result) into motor output, this step may be skipped.

Imitation

Once a robot can coordinate its motors to produce a desired result, the technique of learning by imitation may be used. The robot monitors the performance of another agent and then the robot tries to imitate that agent. It is often a challenge to transform imitation information from a complex scene into a desired motor result for the robot. Note that imitation is a high-level form of cognitive behavior and imitation is not necessarily required in a basic model of embodied animal cognition.

Knowledge acquisition

A more complex learning approach is "autonomous knowledge acquisition": the robot is left to explore the environment on its own. A system of goals and beliefs is typically assumed.

A somewhat more directed mode of exploration can be achieved by "curiosity" algorithms, such as Intelligent Adaptive Curiosity[1][2] or Category-Based Intrinsic Motivation.[3] These algorithms generally involve breaking sensory input into a finite number of categories and assigning some sort of prediction system (such as an Artificial Neural Network) to each. The prediction system keeps track of the error in its predictions over time. Reduction in prediction error is considered learning. The robot then preferentially explores categories in which it is learning (or reducing prediction error) the fastest.

Other architectures

Some researchers in cognitive robotics have tried using architectures such as (ACT-R and Soar (cognitive architecture)) as a basis of their cognitive robotics programs. These highly modular symbol-processing architectures have been used to simulate operator performance and human performance when modeling simplistic and symbolized laboratory data. The idea is to extend these architectures to handle real-world sensory input as that input continuously unfolds through time. What is needed is a way to somehow translate the world into a set of symbols and their relationships.

Questions

Some of the fundamental questions to still be answered in cognitive robotics are:

How much human programming should or can be involved to support the learning processes?

How can one quantify progress? Some of the adopted ways is the reward and punishment. But what kind of reward and what kind of punishment? In humans, when teaching a child for example, the reward would be candy or some encouragement, and the punishment can take many forms. But what is an effective way with robots?[citation needed]

Books

Cognitive Robotics book [4][5] by Hooman Samani,[6] takes a multidisciplinary approach to cover various aspects of cognitive robotics such as artificial intelligence, physical, chemical, philosophical, psychological, social, cultural, and ethical aspects.

Scope

There is growing need for robots that can interact safely with people in everyday situations. These robots have to be able to anticipate the effects of their own actions as well as the actions and needs of the people around them.

collaboration

To achieve this, two streams of research need to merge, one concerned with physical systems specifically designed to interact with unconstrained environments and another focussing on control architectures that explicitly take into account the need to acquire and use experience.

The merging of these two areas has brought about the field of Cognitive Robotics. This is a multi-disciplinary science that draws on research in adaptive robotics as well as cognitive science and artificial intelligence, and often exploits models based on biological cognition.

Cognitive robots achieve their goals by perceiving their environment, paying attention to the events that matter, planning what to do, anticipating the outcome of their actions and the actions of other agents, and learning from the resultant interaction. They deal with the inherent uncertainty of natural environments by continually learning, reasoning, and sharing their knowledge.

A key feature of cognitive robotics is its focus on predictive capabilities to augment immediate sensory-motor experience. Being able to view the world from someone else's perspective, a cognitive robot can anticipate that person's intended actions and needs. This applies both during direct interaction (e.g. a robot assisting a surgeon in theatre) and indirect interaction (e.g. a robot stacking shelves in a busy supermarket).

In cognitive robotics, the robot body is more than just a vehicle for physical manipulation or locomotion: it is a component of the cognitive process. Thus, cognitive robotics is a form of embodied cognition which exploits the robot's physical morphology, kinematics, and dynamics, as well as the environment in which it is operating, to achieve its key characteristic of adaptive anticipatory interaction.

MISSION

The Technical Committee exists to foster links between the fields of robotics, cognitive science, and artificial intelligence. Our goal is to establish and promote the methodologies and tools required to make the field of cognitive robotics industrially and socially relevant.