IntelliMan_WP5_Grasping, Manipulation and Arm-Hand Coordination_T5.1_Data fusion and sensing technology_IEEESENSORS2026_Data

This dataset contains tactile signals acquired using a flexible piezoelectric sensing system integrated into the gripper of a TIAGo robotic platform during controlled object-grasping experiments. The system consists of two flexible sensing arrays, each comprising 16 P(VDF-TrFE) piezoelectric sensing units arranged in a 4 × 4 matrix and embedded in soft silicone sensing caps, providing a total of 32 tactile channels sampled at 500 Hz. The dataset was collected during grasping experiments involving six everyday objects: a glass bottle, a TV remote controller, a half-filled plastic water bottle, a strawberry, a rigid apple, and a semi-soft apple. Each object was grasped 50 times, resulting in a total of 300 trials. Each trial records the tactile responses generated during the grasp–hold–release sequence, with variations in object position and orientation to capture different grasping conditions. The dataset was used to investigate machine learning and deep learning approaches for early-stage object recognition, including support vector machines (SVMs), one-dimensional convolutional neural networks (1D-CNNs), convolutional recurrent neural networks (C-RNNs), and two-dimensional CNNs operating on spatiotemporal tactile heatmaps. Particular attention was devoted to evaluating the influence of the number of sensing units and temporal window size on recognition performance and computational efficiency. The dataset is associated with the following peer-reviewed publication, which provides a detailed description of the sensing system, experimental setup, data acquisition methodology, and embedded object recognition framework: R. Khalifeh, M. Yaacoub, C. Gianoglio, M. Saleh, M. Valle, and Y. Abbass, "Flexible Piezoelectric Tactile Sensing System for Intelligent Object Recognition," IEEE Sensors Journal, vol. 26, no. 8, pp. 12531–12543, 2026. DOI: https://doi.org/10.1109/JSEN.2026.3669666 Users of this dataset are kindly requested to cite the above publication when using the data in their research.