Embedding machine learning methods into the data decoding units may enable the extraction of complex information making the tactile sensing systems intelligent. This paper presents and compares the implementations of a convolutional neural network model for tactile data decoding on various hardware platforms. Experimental results show comparable classification accuracy of 90.88% for Model 3, overcoming similar state-of-the-art solutions in terms of time inference. The proposed implementation achieves a time inference of 1.2 ms while consuming around 900 $\mathsf{\mu }$ J. Such an embedded implementation of intelligent tactile data decoding algorithms enables tactile sensing systems in different application domains such as robotics and prosthetic devices.

Recurrent Neural Networks (RNNs) are mainly designed to deal with sequence prediction problems and they show their effectiveness in processing data originally represented as time series. This paper investigates the time series characteristics of RNNs to classify touch modalities represented as spatio temporal 3D tensor data. Different approaches are followed in order to propose efficient RNN models aimed at tactile data classification. The main idea is to capture long-term dependence from data that can be used to deal with long sequences represented by employing Gated Recurrent Unit (GRU) and Long Short-Term Memory (LSTM) architectures. Moreover, a case specific approach to dataset organization of the 3D tensor data is presented. The target is to provide efficient hardware-friendly touch modality classification approaches suitable for embedded applications. To this end, the proposed work achieves effective performance in terms of hardware complexity by reducing the FLOPS by 99.98% and the memory storage by 98.34%, with respect to the state-of-art solutions on the same benchmark dataset. This directly affects the time latency and energy consumption of the embedded hardware. Besides, the implemented models shows a classification accuracy higher than those state-of-art solutions. Results demonstrate that the proposed computing architecture is scalable showing acceptable complexity when the system is scaled up in terms of input matrix size and number of classes to be recognized.

Embedding Machine Learning enables integrating intelligence in recent application domains such as Internet of Things, portable healthcare systems, and wearable devices. This paper presents an assessment of approximate computing methods at algorithmic, architecture, and circuit levels and draws perspectives for further developments and applications. The main goal is to investigate how approximate computing may reduce the complexity and enable the feasibility of embedded Machine Learning (ML) systems. Though ML is a powerful paradigm for applications in the perceptual domain (i.e. vision, touch, hearing, etc.), their computational complexity is very high and consequently real time operation and ultra-low power are still very challenging objectives. On the other hand, approximate computing has emerged as an effective solution to reduce hardware complexity, time latency and to increase energy efficiency.

Tactile data processing and analysis is still essentially an open challenge. In this framework, we demonstrate a method to achieve touch modality classification using pre-trained convolutional neural networks (CNNs). The 3D tensorial tactile data generated by real human interactions on an electronic skin (E-Skin) are transformed into 2D images. Using a transfer learning approach formalized through a CNN, we address the challenging task of the recognition of the object that was touched by the E-Skin. The feasibility and efficiency of the proposed method are proven using a real tactile dataset outperforming classification results obtained with the same dataset in the literature.

Machine and deep learning algorithms are currently employed for many applications such as computer vision, speech recognition and portable/wearable electronics. Regrettably, machine learning algorithms have high complexity adding more challenges for the implementation of such algorithms on embedded hardware platforms. This paper aims to present an overview about state of the art techniques enabling efficient implementation of Machine and Deep learning (ML/DL) algorithms aiming to improve the energy efficiency. An assessment of the algorithms suitable for embedded implementation is provided, presenting some hardware platforms supporting artificial intelligent systems. On the other hand, we have exploited the choice of implementing ML/DL algorithm on embedded hardware platforms.