GraspLite-Net: A lightweight deep learning method for real-time robotic grasp stability prediction

Abstract

With the rapid advancement of tactile sensing technologies, real-time prediction of robotic grasp stability has become increasingly critical for ensuring safe and reliable operation in automated industrial environments. Achieving accurate real-time predictions of grasp instability is essential for preventing failures during robotic manipulation tasks. Tactile signals serves as a primary source of contact information during grasping and provides detailed insight into the interaction forces and surface dynamics. However, tactile signals often exhibit complex and non-stationary temporal characteristics, posing significant challenges for efficient and robust modeling. To address these challenges while maintaining computational efficiency, we propose GraspLite-Net, a lightweight deep learning-based model that relies solely on tactile time-series data, without requiring visual input. Our proposed GraspLite-Net captures both local and global temporal dependencies while remaining suitable for real-time deployment and comprises three main components: (1) a deformable patch segmentation module that adaptively identifies informative temporal segments; (2) a multi-scale re-parameterized convolutional block for enhanced temporal feature extraction; and (3) an attention-based multiple-instance learning (MIL) pooling module for effective feature aggregation. We conduct extensive experiments on the BiGS dataset and show that GraspLite-Net surpasses baseline methods in both prediction accuracy and inference speed, underscoring its effectiveness and suitability for deployment in real-world robotic grasping systems. We further perform a cross-dataset evaluation on the SnapFitForceProfiles dataset to demonstrate the model’s generalization capability to new robotic scenarios.

Keywords

Grasp stability prediction / Robotic tactile sensing / Lightweight / Deep learning

References

[1] Mohd Aiman Kamarul Bahrin, Mohd Fauzi Othman, Nor Hayati Nor Azli, Muhamad Farihin Talib, Industry 4.0: A review on industrial automation and robotic, J. Teknol. (Sci. Eng.), 78 (6–13), (2016).

[2] Hao Dang, Peter K. Allen, Learning grasp stability, 2012 IEEE International Conference on Robotics and Automation, IEEE, 2012, pp. 2392-2397.

[3] Roland S. Johansson, J. Randall Flanagan, Coding and use of tactile signals from the fingertips in object manipulation tasks, Nature Rev. Neurosci. 10 (5) (2009) 345-359.

[4] Yayu Huang, Dongxuan Fan, Haonan Duan, Dashun Yan, Wen Qi, Jia Sun, Qian Liu, Peng Wang, Human-like dexterous manipulation for anthropomorphic five-fingered hands: A review, Biomim. Intell. Robot. 5 (1) (2025) 100212.

[5] Mark H. Lee, Tactile sensing: new directions, new challenges, Int. J. Robot. Res. 19 (7) (2000) 636-643.

[6] Jiankun Wang, Weinan Chen, Xiao Xiao, Yangxin Xu, Chenming Li, Xiao Jia, Max Q.-H. Meng, A survey of the development of biomimetic intelligence and robotics, Biomim. Intell. Robot. 1 (2021) 100001.

[7] Nicholas Wettels, Veronica J. Santos, Roland S. Johansson, Gerald E. Loeb, Biomimetic tactile sensor array, Adv. Robot. 22 (8) (2008) 829-849.

[8] Yang Gao, Lisa Anne Hendricks, Katherine J. Kuchenbecker, Trevor Darrell, Deep learning for tactile understanding from visual and haptic data, 2016 IEEE International Conference on Robotics and Automation, ICRA, IEEE, 2016, pp. 536-543.

[9] Chao Zhuang, Tianyi Ma, Bokai Xuan, Cheng Chang, Baichuan An, Minghuan Yin, Hao Sun, Deep learning-based semantic segmentation of human features in bath scrubbing robots, Biomim. Intell. Robot. 4 (1) (2024) 100143.

[10] J. Schill, J. Laaksonen, M. Przybylski, V. Kyrki, T. Asfour, R. Dillmann, Learning continuous grasp stability for a humanoid robot hand based on tactile sensing, 2012 4th IEEE RAS & EMBS International Conference on Biomedical Robotics and Biomechatronics, BioRob, IEEE, 2012, pp. 1901-1906.

[11] Johan Tegin, Jan Wikander, Tactile sensing in intelligent robotic manipulation–a review, Ind. Robot.: Int. J. 32 (1) (2005) 64-70.

[12] Hao Dang, Jonathan Weisz, Peter K. Allen, Blind grasping: Stable robotic grasping using tactile feedback and hand kinematics, 2011 Ieee International Conference on Robotics and Automation, IEEE, 2011, pp. 5917-5922.

[13] Yu Zhou, Hao Xia, Dahui Yu, Jiaoyang Cheng, Jichun Li, Outlier detection method based on high-density iteration, Inform. Sci. 662 (2024) 120286.

[14] Jeremy A. Fishel, Gerald E. Loeb, Sensing tactile microvibrations with the BioTac-comparison with human sensitivity, 2012 4th IEEE RAS & EMBS International Conference on Biomedical Robotics and Biomechatronics, BioRob, IEEE, 2012, pp. 1122-1127.

[15] Wenzhen Yuan, Siyuan Dong, Edward H. Adelson, Gelsight: High-resolution robot tactile sensors for estimating geometry and force, Sensors 17 (12) (2017) 2762.

[16] Mike Lambeta, Po-Wei Chou, Stephen Tian, Brian Yang, Benjamin Maloon, Victoria Rose Most, Dave Stroud, Raymond Santos, Ahmad Byagowi, Gregg Kammerer, et al., Digit: A novel design for a low-cost compact high-resolution tactile sensor with application to in-hand manipulation, IEEE Robot. Autom. Lett. 5 (3) (2020) 3838-3845.

[17] Chengliang Liu, Zhengkun Yi, Binhua Huang, Zhenning Zhou, Senlin Fang, Xiaoyu Li, Yupo Zhang, Xinyu Wu, A deep learning method based on triplet network using self-attention for tactile grasp outcomes prediction, IEEE Trans. Instrum. Meas. 72 (2023) 1-14.

[18] Hu Cao, Guang Chen, Zhijun Li, Qian Feng, Jianjie Lin, Alois Knoll, Efficient grasp detection network with Gaussian-based grasp representation for robotic manipulation, IEEE/ASME Trans. Mechatronics 28 (3) (2023) 1384-1394.

[19] Alberto Garcia-Garcia, Brayan S. Zapata-Impata, Sergio Orts-Escolano, Pablo Gil, Jose Garcia-Rodriguez, Tactilegcn: A graph convolutional network for predicting grasp stability with tactile sensors, 2019 International Joint Conference on Neural Networks, IJCNN, IEEE, 2019, pp. 1-8.

[20] Zilin Si, Zirui Zhu, Arpit Agarwal, Stuart Anderson, Wenzhen Yuan, Grasp stability prediction with sim-to-real transfer from tactile sensing, 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS, IEEE, 2022, pp. 7809-7816.

[21] Jingnan Wang, Wenjia Ouyang, Senlin Fang, Yupo Zhang, Xinyu Wu, Zhengkun Yi, TempTrans-MIL: A general approach to enhancing multimodal tactile-driven robotic manipulation classification tasks, IEEE/ASME Trans. Mechatronics, 2025, pp. 1-12.

[22] Yevgen Chebotar, Karol Hausman, Zhe Su, Artem Molchanov, Oliver Kroemer, Gaurav Sukhatme, Stefan Schaal, Bigs: Biotac grasp stability dataset, ICRA 2016 Workshop on Grasping and Manipulation Datasets, vol. 21, (2016), p. 23.

[23] Navid Mohammadi Foumani, Lynn Miller, Chang Wei Tan, Geoffrey I. Webb, Germain Forestier, Mahsa Salehi, Deep learning for time series classification and extrinsic regression: A current survey, ACM Comput. Surv. 56 (9) (2024) 1-45.

[24] Jingnan Wang, Zhenning Zhou, Yunduan Cui, Jian Cui, Longhui Qin, Tiantian Xu, Zhengkun Yi, Xinyu Wu, Dual autoencoder-based joint learning: Enhancing depth classification of hard inclusions in soft tissue for robotic palpation, IEEE/ASME Trans. Mechatronics, 2025, pp. 1-12.

[25] Jingnan Wang, Xiaoyu Li, Yunduan Cui, Ke Mai, Yiru Wang, Mianzhi Song, Can Wang, Zhengkun Yi, Xinyu Wu, Soft sensor-based deep temporal-graph convolutional network for applications in human motion tracking, IEEE Sensors J. 24 (14) (2024) 23117-23128.

[26] Zhifei Shen, Zhiyong Jiang, Jingwang Zhang, Jun Wu, Qiuguo Zhu, Learning-based robot assembly method for peg insertion tasks on inclined hole using time-series force information, Biomim. Intell. Robot. 5 (1) (2025) 100209.

[27] David Campos, Miao Zhang, Bin Yang, Tung Kieu, Chenjuan Guo, Christian S. Jensen, Lightts: Lightweight time series classification with adaptive ensemble distillation, Proc. ACM Manag. Data 1 (2) (2023) 1-27.

[28] Zhichen Lai, Dalin Zhang, Huan Li, Christian S. Jensen, Hua Lu, Yan Zhao, Lightcts: A lightweight framework for correlated time series forecasting, Proc. ACM Manag. Data 1 (2) (2023) 1-26.

[29] Lei Chen, Jiajun Tang, Ying Zou, Xuxin Liu, Xingquan Xie, Guangyang Deng, Lightweight and fast time-series anomaly detection via point-level and sequence-level reconstruction discrepancy, IEEE Trans. Neural Netw. Learn. Syst., (2025).

[30] Maximilian Ilse, Jakub Tomczak, Max Welling, Attention-based deep multiple instance learning, International Conference on Machine Learning, PMLR, 2018, pp. 2127-2136.

[31] Zhengkun Yi, Tiantian Xu, Wanfeng Shang, Weimin Li, Xinyu Wu, Genetic algorithm-based ensemble hybrid sparse ELM for grasp stability recognition with multimodal tactile signals, IEEE Trans. Ind. Electron. 70 (3) (2022) 2790-2799.

[32] Yanping Chen, Bing Hu, Eamonn Keogh, Gustavo E.A.P.A. Batista, Dtw-d: time series semi-supervised learning from a single example, in: Proceedings of the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2013, pp. 383-391.

[33] Fazle Karim, Somshubra Majumdar, Houshang Darabi, Samuel Harford, Multivariate LSTM-FCNs for time series classification, Neural Netw. 116 (2019) 237-245.

[34] Donghao Luo, Xue Wang, Moderntcn: A modern pure convolution structure for general time series analysis, in: The Twelfth International Conference on Learning Representations, 2024.

[35] Wensi Tang, Guodong Long, Lu Liu, Tianyi Zhou, Michael Blumenstein, Jing Jiang, Omni-Scale CNNs: a simple and effective kernel size configuration for time series classification, in: International Conference on Learning Representations, 2021.

[36] Huaiyuan Liu, Donghua Yang, Xianzhang Liu, Xinglei Chen, Zhiyu Liang, Hongzhi Wang, Yong Cui, Jun Gu, Todynet: temporal dynamic graph neural network for multivariate time series classification, Inform. Sci., (2024), 120914.

[37] Tingting Mi, Dashun Que, Senlin Fang, Zhenning Zhou, Chaoxiang Ye, Chengliang Liu, Zhengkun Yi, Xinyu Wu, Tactile Grasp Stability Classification Based on Graph Convolutional Networks, in: 2021 IEEE International Conference on Real-Time Computing and Robotics, RCAR, 2021, pp. 875-880.

[38] Stefanos Doltsinis, Marios Krestenitis, Zoe Doulgeri, A machine learning framework for real-time identification of successful snap-fit assemblies, IEEE Trans. Autom. Sci. Eng. 17 (1) (2019) 513-523.

[39] Imen Boulnemour, Bachir Boucheham, Abdelmadjid Lahreche, On enhancing the accuracy of nearest neighbour time series classifier using improved shape exchange algorithm, Int. J. Inform. Appl. Math. 4 (1) (2021) 15-27.