no image available

Dr Paul Baxter

Research Fellow in Human-Robot Interaction


  • PhD Cybernetic Intelligence (Developmental Cognitive Robotics), University of Reading (U.K.)
  • MEng Electronic Engineering, University of Nottingham (U.K.)

Research interests

My primary interest lies with the principle of memory-based cognition, particularly as applied to cognitive robotics. This principle states that cognition is the manipulation and utilisation of memory - the two cannot be dissociated. This has some very interesting implications for cognitive robotics and cognitive architecture implementations, and it is these that I am interested in investigating. For instance, it commits you to a common substrate for perception, action, cognition, learning (which may be regarded as the adaptation of memory), etc, and only really makes sense in the framework of constructivism (the process of development through experience-based construction of 'control' and 'knowledge' structures), through ongoing environmental interaction. Extending these ideas to social interaction in the context of human-robot interaction is becoming an equally important aspect of my research into these ideas of cognitive robotics.

My current research is conducted as part of the EU FP7 DREAM project at Plymouth University (Plymouth PI is Prof. Tony Belpaeme), where I am focusing on the cognitive and behavioural aspects of robot control to support robot-assisted therapy for autistic children. The idea here is that robots will be able to interact autonomously with autistic children for both diagnosis and therapy, though under the supervision of a therapist (supervised autonomy).

Previously, I worked as part of the EU ALIZ-E project (coordinated by Prof. Tony Belpaeme), researching memory systems to support long-term human-robot interactions. I attempted to incorporate aspects (wherever practicable) of the memory-based cognition principle to this work, resulting in a proposed departure from current models of memory for computational implementation. For example, it led to a far more active role for 'memory' than normally ascribed to it, being centrally involved in the coordination and generation of ongoing behaviour, rather than the memory system acting as a passive adjunct to the cognitive processing. Apart from the consequences for the use of memory, it also has profound implications for the nature of cognitive architecture, which I have also explored, and continue to explore. Furthermore, this work in human-robot interaction emphasises the point that the environment in which cognitive robotic systems interact is a social one, a principle I continue to apply.

In addition to this, I cultivate an interest in a wide range of biological cognition and cognitive robotic-related fields, including (and here is the compulsory non-exhaustive list of keywords for a page like this, in no particular order), morphological computation (particularly passive dynamic walkers), theories of embodiment (embodied cognition), motivation/value/emotion systems, cognitive development, sensorimotor coordination, grounding (including, but not limited to, symbol grounding) and consciousness.