The main objective is to undertake original research that will lead to a biomimetic tactile sensor suitable for incorporating in the fingerpad of an artificial finger. We hope to develop sufficient scientific understanding of the operational characteristics of the human mechanoreceptors, the factors that influence their recruitment and discharge patterns and the human neural coding of taction. We shall evaluate our success by the ability of the artificial finger to assess complex tactile stimuli involving different textures and surface qualities such as stickiness, not just spatial acuity, geometric form, surface topography and vibration, which the artificial finger should also possess. The milestones on the way to this objective are as follows:

1. To identify a range of relevant textures and surface qualities as the basis for the tactile stimuli that will be characterised

2. To measure the peripheral and brain response to the tactial stimuli and characterise these response in terms of global and time series statistics

3. To develop a validated finite element virtual model of the tactile stimulation of a fingerpad that incorporates time-resolved neuromechanical coupling

4. To model the interrelationships between:

• The interaction mechanics of tactile stimuli
• The real and computer simulated peripheral neural responses
• The brain imaging data
• The effects of peripheral and brain lesions
• The pschophysical assessment of the stimuli Tactile discrimination algorithms will be developed for this purpose.

5. To develop a biomimetic sensor and incorporate it into an artificial finger with the output signals processes to provide tactile assessment and motor feedback Links Original research

To back up these objectives, we shall undertake original research as follows:

1. The psychophysical effects of the tactile stimuli will be characterised in physical detail including the measurement of the global and local, normal and tangential, statistic and dynamic contact stresses. Such characterisation is usually limited to primitive measures in psychophysical studies.

2. The neurophysiological work will be undertaken at a number of levels of analysis using complex tactile stimuli in a manner that has not previously been attempted.

3. The finite element simulation of the fingerpad will be based on advanced optimisation procedures for obtaining the material models from improved mechanical measurements of the fingerpad. It will be highly spatial resolved with dynamic neuromechanical coupling to simulate real tactile experiences.

4. The availability of such a wide range of detailed data from different techniques and from unstudied stimuli is a unique resource for understanding taction. The application of artificial recurrent neural networks to neurophysiological data to account for the sequential nature of taction has not been attempted previously and this will provide novel tactile discrimination software.

5. Computer-aided gradient based optimisation techniques will be used to design the individual NEMS force transducers and also the biomimetic array by exploiting the finite element model of fingerpad taction. Novel approaches to biomimetic packaging materials based on tissue engineering will be used to emulate the human fingerpad.