Neuroshoes is part of the Neurosuit project, which aims to create a new form of embodiment by combining the user’s sensorimotor coupling learning, generating and transferring qualia to others, and integrating the autonomy of both the user and the system. The neuroshoes project explores new possibilities for shoes from the perspective of artificial life, and embodied mind.

The Neuroshoes prototype incorporates various sensors, including foot pressure sensors, IMUs, and microphones, as well as actuators like vibration speakers, and audio speakers. The texture of the walking sounds generated by the vibration speakers can create the sensation of different surfaces underfoot. For example, an office hallway can feel like a desert or snowy mountain, or the heavy footsteps of a robot can make the user naturally slow down or feel a sense of weight, inducing various perceptual changes through altered foot sensations.

The neural network embedded in the Neuroshoes prototype learns a predictive model of the user’s walking patterns from sensor inputs. Simultaneously, it uses prediction errors as rewards to learn to adjust the timing of the walking sounds produced by the vibration speakers. This approach is expected to elicit unpredictable walking patterns. By deliberately disrupting the typically unconscious experience of walking, the prototype attempts to transform it into a lively and invigorating experience, reminiscent of learning to walk for the first time.

The Neuroshoes project will continue to explore new possibilities for shoes that provide vivid experiences by not only enhancing sensing capabilities but also adding various types of actuation.