Mark Witkowski's Home Page

Keywords: Cognitive Robotics, Dynamic Expectancy Theory, Anticipatory Learning, Simulation of Adaptive Behaviour, Biologically Inspired Robotics, Action Selection

Circuits and Systems Group,
Department of Electrical and Electronic Engineering,
Imperial College London,
Room 913c
Exhibition Road,
London, SW7 2BT
United Kingdom.
Tel: +44 (0)20 7594 6205; Fax: +44 (0)20 7581 4419
m.witkowski@imperial.ac.uk

New

"London Across Time" (Według Czasu Londyńskiego): an exhibition of stereo (3D) photographs at the Warsaw Fotoplastikon, showing how London has changed since the 1960s. This exhibition of 48 stereographs combines and contrasts stereo photographs by my father Stanisław Witkowski, taken in the late 1960s, with modern stereo photographs by myself to show a changing and changed city across the two generations. The exhibition runs from 21st April 2011, see the poster here (in English, in Polish). More details here.

The Warsaw Fotoplastikon (Stereoscope) is an authentic late 19th century viewing machine for stereoscopic images, fully functional and operating today on its original site in central Warsaw. The Fotoplastikon is part exhibition space, part museum and part tourist attraction, and holds regular exhibitions of modern and historical stereo photographs. Well worth a visit whenever in Warsaw!  

TAROS-2011, August 31st to September 2nd, Sheffield University, UK. The 12th Towards Autonomous Robotics Systems Conference includes an Industry Day (2nd September) to foster and strengthen links between academia and industry.

Research Interests

My on-going "background" research goal is to consolidate the work described below on various aspects of perception, attention, awareness, planning, learning, behaviour generation and reasoning into a unified model for application to robotic controller design. I have taken as one of my conceptual starting points Igor Aleksander's "Five Axioms" (sense of presence, imagination, attention, planning and emotion), while recognising that they will not be sufficient to realise human level machine awareness and activity, they are each almost certainly necessary to achieve it. Whereas Aleksander's axioms are concerned with high-level description, my primary concern is with underlying principles of operation necessary to support these diverse observables. Unsurprisingly, I'm not expecting immediate results, but the journey is filled with endless fascination. 

In the meanwhile I am collaborating with Prof. Bob Spence and James Mardell in the Circuits and Systems Group at Imperial College London on aspects of Human-Computer interface. We are looking at visual inspection for simulated search and rescue tasks and aspects of the human attentional system. We've also been looking at eyegaze contingent control of large image space, with Nick Adams, who developed and thoroughly investigated a very elegant user interface strategy for "hands-free" navigation of Google Earth just by looking at the screen. This work was awarded the HCI 08 Award for International Excellence. You can find this and a list of selected publications here. 

I am also a Visiting Fellow at the Bristol Robotics Laboratory, and I am working with the team there to investigate aspects of Human-Robot Interaction in the context of safe robot operation when in close proximity to humans.

I have recently worked on understanding intentionality and threat space in a tactical environment with Prof. Erol Gelenbe's Intelligent Systems and Networks group and with Prof. Panos Louvieris (Brunel University) on decision support methodologies in an adversarial context, both part of the UK Defence Technology Centre/Data Information Fusion Initiative (DTC). 

I have worked with Murray Shanahan in his Cognitive Robotics group, with Dr. David Randell on formal representations for robot planning, sensory interpretation, perception and attention. Our most recent, completed, EPSRC project "Abductive Robot Perception: Modelling Granularity and Attention in Euclidean Representational Space" takes an abductive (reasoning from observations to postulated causes) reasoning approach to active perception for robotics. Key results from this work include a formal view of the perception process and a consolidation of multiple aspects of attention and awareness in this formalised view. Perception is taken as a combination of bottom-up sense driven and top-down hypothesis driven activities. Features extracted from the robot sensors indicate explanatory hypotheses in the bottom-up phase. Subsequently, the deductive consequences of each of (the many, possibly competing) hypotheses are used to disambiguate between alternative interpretations of the sense data to provide an explanation of that data. This project used our upper torso humanoid robot Ludwig. Our previous projects in the Imperial College Cognitive Robotics series have been Spatial Reasoning and Perception in a Humanoid Robot, Cognitive Robotics II and Cognitive Robotics I, which investigated the use of combines Shanahan's existing Event Calculus formalism in an abductive framework to investigate a range of topics in robot sensor data assimilation, planning and action selection. 

My on-going research interests include the simulation and understanding of natural Behaviour through anticipatory learning methods and associated action selection strategies. My latest paper on these issues "An Action Selection Calculus" is in the Adaptive Behavior Journal. These are a continuation of earlier work on the "Dynamic Expectancy Theory", which integrates innate reactive (pre-programmed or evolved) behaviours with goal-directed learned behaviour to address the action selection problem. Learning in the Dynamic Expectancy Model is accomplished by an ingenious variant of the reinforcement learning method in which expectancies about the environment ("micro-hypotheses") are created and corroborated (by "micro-experiments") by a predictive method which dispenses with the need for the provision of external reward to drive learning. Motivation and learning are separated and learning may take place latently, independently of any known or assumed goal. The system generates reactive goal-directed behaviours by formulating a "dynamic policy map" from stored micro-hypotheses when goals arise.

Previous research and commercial interests have been in robotics, industrial local area networking, expert systems for dentistry and system configuration, and methodologies for manufacturing systems and product design.

I am also a director of AIQ Limited, which offers design and consultancy services in these and related areas.

TAROS-11 will be in Sheffield 31st August to 2nd September, with a special industry day on the 2nd of September. I was local organisation chair for Towards Autonomous Robotics Conference (TAROS-05) at Imperial College in September 2005 (on-line archive), chair of the programme committee for TAROS-06 (proceedings, large file) held at Surrey University in September 2006, and conference co-chair for TAROS-07 held in Aberystwyth in September 2007 and TAROS-08 in Edinburgh, 1-3 September 2008. TAROS-09 in Ulster 31 August to 2 September 2009. TAROS-10 was in Plymouth.

Please contact me by e-mail at the Department.

Papers and Publications

Thesis, full text in pdf and postscript

Some Photos of Past Robot Projects