Teresa Escrig

News and oppinion about Cognitive AI & Robotics

Cognitive Robots includes Common-Sense Knowledge and Reasoning into their Robotics and Computer Vision solutions


Representation, reasoning and learning are the basic principles of human intelligence. The emulation of human intelligence has been the aim of Artificial Intelligence since its origins in 1956.

In fact, converting raw data into information (data in the context of other data) and hence into knowledge (information in the context of other information), is critical for understanding activities, behaviors, and in general the world we try to model. Both in the Robotics and the Computer Vision areas we try to model the real world where the humans are operating.

The type of knowledge that Robotics and Computer Vision need to obtain is Common Sense Knowledge. Contra intuitively, common sense knowledge is more difficult to model than expert knowledge, which can be quite easily modeled by expert systems (a more or less closed research area since the 70s).

Both in Robotics and Computer Vision areas, Probabilistic and Bayesian models have historically been used as the way to represent, reason and learn from the world. These methods have provided very good initial results. The problem is that they have never been scalable. That is why there is no commercial intelligent robot that has the full ability to serve people yet. Although there exist many preliminary solutions including artificial vision, the percentage of false positives or negatives are still too high to consider it as completely reliable, and therefore artificial vision is still an open research area.

The problems detected in the probabilistic approaches have been twofold:

  • It is a brute force method with high computational cost.
  • There is no use of common sense or any other cognitive approach to make sense of the numbers, therefore the improvements are limited.

We need “more intelligence” to be able to automate common sense.

Qualitative Representation and Reasoning, on the other hand, have been demonstrated to be a much better approach to model common sense information, by transforming  uncertainty and incomplete data into knowledge. Highly promising results have been obtained in the area of Robotics for autonomous Map Building, Localization and Navigation [1,2,3,4].

This is the approach that Cognitive Robots has followed since 1992.

20 years of research (at the University Jaume I, Spain) and development are now included in the most advanced intelligent system that can be incorporated into physical and virtual devices to increase their intelligent, in an immense variety of applications.

Cognitive Robots is now actively looking for commercial partnerships and investment capital to bring the company into the next level. If you are interested, please contact investors@c-robots.com.


1. Escrig, M.T., Peris, J.C, “The use of a reasoning process to solve the almost SLAM problem at the Robocup legged league”, Catalonian Conference on Artificial Intelligence, CCIA’05, 2005.

2. Peris, J.C, Escrig, M.T., “Cognitive Maps for mobile Robot Navigation: A Hybrid Representation Using Reference Systems”, 19th International Workshop on Qualitative Reasoning, Graz, Austria, pp. 179-185, ISBN 3-9502019-0-4, Graz, Austria, 2005.


4. Z. Falomir, Ll. Museros, L. Gonzalez-Abril, M.T. Escrig, J.A. Ortega, “A model for the qualitative description of images based on visual and spatial features”, Computer Vision and Image Understanding 116, 698–714, 2012.


5 Responses to 'Cognitive Robots includes Common-Sense Knowledge and Reasoning into their Robotics and Computer Vision solutions'

Subscribe to comments with RSS or TrackBack to 'Cognitive Robots includes Common-Sense Knowledge and Reasoning into their Robotics and Computer Vision solutions'.

  1. Dear Teresa,

    I don’t see why Qualitative Representation and Reasoning would be a “much better approach” than the probabilistic one. I think instead the solution is to integrate both qualitative and probabilistic methods to create more powerful solutions for solving problems in mobile robotics.

    Best regards,


    PS: you might be interested in a couple of our recent publications:

    N. Bellotto, Robot Control based on Qualitative Representation of Human Trajectories, AAAI Spring Symposium “Designing Intelligent Robots: Reintegrating AI”, TR SS-12-02, 2012.

    M. Hanheide, A. Peters and N. Bellotto, Analysis of Human-Robot Spatial Behaviour applying a Qualitative Trajectory Calculus, Proceedings of the IEEE International Symposium on Robot and Human Interactive Communication (Ro-Man 2012) (accepted)

    Nicola Bellotto

    27 Jul 12 at 9:33 am

  2. Hi Nicola,

    You are right with your statement.
    I get in trouble when I write in absolute terms. I need to take care of my way of expression. Thank you for showing me that.

    In some aspects of robotics, probabilistic solutions can be useful. And integration of the best of both parts, is the optimum.

    We were working in qualitative trajectories and made significant contributions a few years ago. I love the subject. Could you, please, send your articles to my email?

    Thanks again Nicola!

    Teresa Escrig

    31 Jul 12 at 12:24 am

  3. “In some aspects of robotics, probabilistic solutions can be useful. And integration of the best of both parts, is the optimum.”

    Research to find solutions to achieve such integration (even if not always formulated as such) was very active in the 90s and early 2000. Numerous robots hybrid architectures have been proposed, with the huge drawback of being almost always robot dependent and/or task dependent. There were also not open-source and I suspect extremely difficult to reuse. Since then, this domain of research has evolved in the direction of research on middleware, with a lot of success (Player, ROS …).

    But the issue a software that would give easy/reusable/flexible solutions for simultaneous integration of deliberative cognition (reasoning based on representative knowledge) and reaction (often relying on probabilistic methods) has been (almost) forgotten. Does the work of Cognitive Robots have anything to do with this, or is it something different ?



    3 Sep 12 at 2:15 pm

  4. Hi Vincent,

    Thank you for your comment.
    Cognitive Brain for Service Robotics, CR-B100, has been developed in ROS, therefore the core of the system (representation and reasoning) is hardware independent, reusable.

    We do not use probabilistic methods but qualitative representation and reasoning methods instead, which naturally deal with uncertainty, are much faster and cognitive. We also capture quantitative data in our system just in case we need it, but we are not dependent on quantitative data, therefore we do not carry on errors.

    Does this answer your question?

    Kind regards,

    Teresa Escrig

    4 Sep 12 at 5:47 pm

  5. Hi Ashley,

    Yes, we are considering everything.
    We do not see Kickstarter as the best crow-funding platform for us because our product doesn’t cost hundreds of dollars.
    Do you know of a platform that can fit better our product?

    Thank you for caring!


    Teresa Escrig

    25 Oct 12 at 2:58 pm

Leave a Reply