This technology offer presents a deep learning-enabled smart mat, which is based on the triboelectric mechanism. This deep learning-enabled smart mat is achieved through the integration of a minimal-electrode-output triboelectric floor mat array with advanced deep learning (DL)-based data analytics. By walking/stepping over the smart mat, the unique fingerprint-like output signal can be captured by the electrode array. With the integrated deep learning-based data analytics, identity information associated with walking gait patterns can be extracted from the output signals using the convolutional neural network (CNN) model.
The smart mat can be applied in automation control, such as lighting and air conditioning and fall detection. The smart mat can also be adopted for activity monitoring (e.g., walking, running, exercising) and potential energy harvesting from our daily activities.
The deep learning-enabled smart mats are fabricated by screen printing, exhibiting the merits of cost-effectiveness, high scalability, and self-sustainability in large-area applications. A distinct electrode pattern with varying coverage rates is designed for each deep learning-enabled smart mat, mimicking the unique identification of the QR (quick response) code system. Hence, after the parallel connection in an interval scheme, minimal two-electrode outputs with distinguishable and stable characteristics for the whole deep learning-enabled smart mat array can be achieved.