Abstract: Robots need more intelligence to complete cognitive tasks in home environments. In this paper, we present a new cloud-assisted cognition adaptation mechanism for home service robots, which learns new knowledge from other robots. In this mechanism, a change detection approach is implemented in the robot to detect changes in the user‘s home environment and trigger the adaptation procedure that adapts the robot‘s local customized model to the environmental changes, while the adaptation is achieved by transferring knowledge from the global cloud model to the local model through model fusion. First, three different model fusion methods are proposed to carry out the adaptation procedure, and two key factors of the fusion methods are emphasized. Second, the most suitable model fusion method and its settings for the cloud–robot knowledge transfer are determined. Third, we carry out a case study of learning in a changing home environment, and the experimental results verify the efficiency and effectiveness of our solutions. The experimental results lead us to propose an empirical guideline of model fusion for the cloud–robot knowledge transfer.

面向变化用户家居环境的服务机器人云辅助认知适应

[1]Arumugam R, Enti VR, Liu BB, et al., 2010. DAvinCi: a cloud computing framework for service robots. IEEE Int Conf on Robotics and Automation, p.3084-3089. doi: 10.1109/ROBOT.2010.5509469

[2]Baena-García M, del Campo-Ávila J, Fidalgo R, et al., 2006. Early drift detection method. Proc 4^th Int Workshop on Knowledge Discovery from Data Streams, p.77-86.

[3]Disabato S, Roveri M, 2019. Learning convolutional neural networks in presence of concept drift. Int Joint Conf on Neural Networks, p.1-8. doi: 10.1109/IJCNN.2019.8851731

[4]Ditzler G, Roveri M, Alippi C, et al., 2015. Learning in nonstationary environments: a survey. IEEE Comput Intell Mag, 10(4):12-25. doi: 10.1109/MCI.2015.2471196

[5]Fachantidis N, Dimitriou AG, Pliasa S, et al., 2018. Android OS mobile technologies meets robotics for expandable, exchangeable, reconfigurable, educational, stem-enhancing, socializing robot. Interactive Mobile Communication Technologies and Learning, p.487-497. doi: 10.1007/978-3-319-75175-7_48

[6]Gama J, Medas P, Castillo G, et al., 2004. Learning with drift detection. Proc 17^th Brazilian Symp on Artificial Intelligence, p.286-295. doi: 10.1007/978-3-540-28645-5_29

[7]Gouveia BD, Portugal D, Silva DC, et al., 2015. Computation sharing in distributed robotic systems: a case study on SLAM. IEEE Trans Autom Sci Eng, 12(2):410-422. doi: 10.1109/TASE.2014.2357216

[8]Graf B, Staab H, 2009. Service robots and automation for the disabled/limited. In: Nof SY (Ed.), Springer Handbook of Automation. Springer Berlin Heidelberg, p.1485-1502. doi: 10.1007/978-3-540-78831-7_84

[9]Huaimin W, Bo D, Xu J, 2018. Cloud robotics: a distributed computing view. In: Jones C, Wang J, Zhan NJ (Eds.), Symp on Real-Time and Hybrid Systems. Springer, Cham, p.231-245. doi: 10.1007/978-3-030-01461-2_12

[10]Hunziker D, Gajamohan M, Waibel M, et al., 2013. Rapyuta: the RoboEarth cloud engine. IEEE Int Conf on Robotics and Automation, p.438-444. doi: 10.1109/ICRA.2013.6630612

[11]Kuncheva LI, 2013. Change detection in streaming multivariate data using likelihood detectors. IEEE Trans Knowl Data Eng, 25(5):1175-1180. doi: 10.1109/TKDE.2011.226

[12]LeCun Y, Bengio Y, Hinton G, 2015. Deep learning. Nature, 521(7553):436-444. doi: 10.1038/nature14539

[13]McMahan HB, Moore E, Ramage D, et al., 2017. Communication-efficient learning of deep networks from decentralized data. https://arxiv.org/abs/1602.05629

[14]Page ES, 1954. Continuous inspection schemes. Biometrika, 41(1-2):100-115. doi: 10.1093/biomet/41.1-2.100

[15]Pedregosa F, Varoquaux G, Gramfort A, et al., 2011. Scikit-learn: machine learning in Python. J Mach Learn Res, 12:2825-2830.

[16]Secker J, Hill R, Villeneau L, et al., 2003. Promoting independence: but promoting what and how? Ageing Soc, 23(3):375-391. doi: 10.1017/S0144686X03001193

[17]Shuai W, Chen XP, 2019. KeJia: towards an autonomous service robot with tolerance of unexpected environmental changes. Front Inform Technol Electron Eng, 20(3):307-317. doi: 10.1631/FITEE.1900096

[18]Siciliano B, Khatib O, 2016. Springer Handbook of Robotics. Springer Berlin Heidelberg, Germany.

[19]Simonyan K, Zisserman A, 2014. Very deep convolutional networks for large-scale image recognition. https://arxiv.org/abs/1409.1556

[20]TensorFlow, 2020. TensorFlow-Slim Image Classification Model Library. https://github.com/tensorflow/models/tree/master/research/slim

[21]United Nations, 2020. Ageing. https://www.un.org/en/sections/issues-depth/ageing/

[22]Wang Q, Zhang S, Sheng W, et al., 2021. Multi-style learning for adaptation of perception intelligence in home service robots. Patt Recogn Lett, 151:243-251.

[23]Yang Z, Al-Dahidi S, Baraldi P, et al., 2020. A novel concept drift detection method for incremental learning in nonstationary environments. IEEE Trans Neur Netw Learn Syst, 31(1):309-320. doi: 10.1109/TNNLS.2019.2900956

[24]Yoon J, Jeong W, Lee G, et al., 2020. Federated continual learning with adaptive parameter communication. https://arxiv.org/abs/2003.03196v1

[25]Yosinski J, Clune J, Bengio Y, et al., 2014. How transferable are features in deep neural networks? https://arxiv.org/abs/1411.1792v1

[26]Žliobaitė I, 2010. Learning under concept drift: an overview. https://arxiv.org/abs/1010.4784

[27]Žliobaitė I, Pechenizkiy M, Gama J, 2016. An overview of concept drift applications. In: Japkowicz N, Stefanowski J (Eds.), Big Data Analysis: New Algorithms for a New Society. Springer, Cham, p.91-114. doi: 10.1007/978-3-319-26989-4_4

[28]Zweigle O, van de Molengraft R, d’Andrea R, et al., 2009. RoboEarth: connecting robots worldwide. Proc 2^nd Int Conf on Interaction Sciences: Information Technology, p.184-191. doi: 10.1145/1655925.1655958