識別流動關係對於理解人類行為和社會互動是非常重要的。隨著穿戴式裝置、5G 和計算能力的蓬勃發展，我們的生活中存在大量可利用的資訊。在本文中，我們開發了一個框架來識別移動群體的軌跡和鄰近關係。軌跡關係分為同伴、領導者、追隨者、獨立等四種關係。鄰近關係問題將領導者和追隨者之間的距離分為 1m、2m、3m 和 5m。該框架混合視覺和無線訊號以適應不同的環境並訓練機器學習模型來識別任務。首先，我們利用雙視角視覺哈希(Dual-view Visual Hash)來提取特徵並計算兩個第一人稱視角影片之間的相似度。其次，我們開發了無線哈希(Wireless Hash)而非無線指紋來捕獲空間特徵。形式上，該框架將這兩種特徵融合為一個混合特徵。這兩種特徵都是基於哈希算法生成的。哈希算法克服了計算複雜性並獲得了更高的性能。我們進行現實生活中的實驗和腳本化實驗。實驗結果表明，我們提出的方法在軌跡任務中達到了 95% 的準確率，在接近識別中達到了 90%。此外，該方法也優於現有方法。

Recognition of mobility relationships is important for understanding human behavior and social interaction. With the flourishing of wearable devices, 5G and computation, there exist the massive amount of information surrounding our life. In this paper, we develop a framework to recognize trajectory and proximity relationships of mobility groups. The trajectory relationships divide into four level, including companion, leader, follower, independent. The proximity tasks classify the distance between leaders and followers for 1m, 2m, 3m and 5m. The framework jointly incorporates visual and wireless information to adapt to the different environments and train the ma- chine learning model to recognize the tasks. First, we exploit dual-view visual hash to extract the feature and compute the similarity between two first-person view video. Second, we develop a wireless hash instead of wireless fingerprints to capture spatial feature. Formally, the framework fuse the two information to a hybrid feature. Both of the feature are generated based on hash algorithms. The hash algorithms conquer the computational complexity and achieve higher performance. We conduct the real-life experiments and scripted experiments. The experimental result shows that our proposed method achieves an accuracy of 95% in trajectory tasks and an 90% in proximity recognition. In addition, the method also outperforms the existing approach.

[1] H. A. Naji, C. Wu, H. Zhang, and L. Li, “Towards understanding the impact of human mobility patterns on taxi drivers’ income based on gps data: A case study in wuhan—china,” in 2017 4th International Conference on Transportation Information and Safety (ICTIS). IEEE, 2017, pp. 1152–1160.
[2] J. Shen, J. Cao, X. Liu, and S. Tang, “Snow: detecting shopping groups using wifi,” IEEE Internet of Things Journal, vol. 5, no. 5, pp. 3908–3917, 2018.
[3] Y. Chon, S. Kim, S. Lee, D. Kim, Y. Kim, and H. Cha, “Sensing wifi packets in the air: practicality and implications in urban mobility monitoring,” in Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing, 2014, pp. 189–200.
[4] X. Andrade, F. Layedra, C. Vaca, and E. Cruz, “Risc: Quantifying change after natural disasters to estimate infrastructure damage with mobile phone data,” in 2018 IEEE International Conference on Big Data (Big Data). IEEE, 2018, pp. 3383–3391.
[5] J.-Y. Tsai and S.-I. Sou, “Proximity detection for social distancing based on wireless signal strength time series,” 2020.
[6] T.-H. Chen, S.-I. Sou, and Y. Lee, “Witrack: Human-to-human mobility relationship tracking in indoor environments based on spatio-temporal wireless signal strength,” in 2019 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech). IEEE, 2019, pp. 788–795.
[7] R. San-Segundo, J. D. Echeverry-Correa, C. Salamea, and J. M. Pardo, “Human activity monitoring based on hidden markov models using a smartphone,” IEEE Instrumentation & Measurement Magazine, vol. 19, no. 6, pp. 27–31, 2016.
[8] J. Wang, Y. Chen, S. Hao, X. Peng, and L. Hu, “Deep learning for sensor-based activity recognition: A survey,” Pattern Recognition Letters, vol. 119, pp. 3–11, 2019.
[9] S. Yao, S. Hu, Y. Zhao, A. Zhang, and T. Abdelzaher, “Deepsense: A unified deep learning framework for time-series mobile sensing data processing,” in Proceedings of the 26th International Conference on World Wide Web, 2017, pp. 351–360.
[10] Q. Wang, M. Chen, F. Nie, and X. Li, “Detecting coherent groups in crowd scenes by multiview clustering,” IEEE transactions on pattern analysis and machine intelligence, vol. 42, no. 1, pp. 46–58, 2018.
[11] A. Roy, N. Biswas, S. K. Saha, and B. Chanda, “Classification of moving crowd based on motion pattern,” in 2019 IEEE Region 10 Symposium (TENSYMP). IEEE, 2019, pp. 102–107.
[12] G. Solmaz and F.-J. Wu, “Together or alone: Detecting group mobility with wireless fingerprints,” in 2017 IEEE International Conference on Communications (ICC). IEEE, 2017, pp. 1–7.
[13] J. Shen, J. Cao, and X. Liu, “Bag: behavior-aware group detection in crowded urban spaces using wifi probes,” IEEE Transactions on Mobile Computing, 2020.
[14] H. Du, Z. Yu, F. Yi, Z. Wang, Q. Han, and B. Guo, “Recognition of group mobility level and group structure with mobile devices,” IEEE Transactions on Mobile Computing, vol. 17, no. 4, pp. 884–897, 2017.
[15] K. Katevas, K. Hänsel, R. Clegg, I. Leontiadis, H. Haddadi, and L. Tokarchuk, “Finding dory in the crowd: Detecting social interactions using multi-modal mobile sensing,” in Proceedings of the 1st Workshop on Machine Learning on Edge in Sensor Systems, 2019, pp. 37–42.
[16] J. Li, Y. Wong, Q. Zhao, and M. S. Kankanhalli, “Dual-glance model for deciphering social relationships,” in Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 2650–2659.
[17] S. Alletto, G. Serra, S. Calderara, F. Solera, and R. Cucchiara, “From ego to nosvision: Detecting social relationships in first-person views,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2014, pp. 580–585.
[18] A. Bochkovskiy, C.-Y. Wang, and H.-Y. M. Liao, “Yolov4: Optimal speed and accuracy of object detection,” arXiv preprint arXiv:2004.10934, 2020.
[19] C. Zauner, “Implementation and benchmarking of perceptual image hash functions,” 2010.
[20] J.-P. Aumasson, S. Neves, Z. Wilcox-O’Hearn, and C. Winnerlein, “Blake2: simpler, smaller, fast as md5,” in International Conference on Applied Cryptography and Network Security. Springer, 2013, pp. 119–135.
[21] G. Solmaz, J. Fürst, S. Aytaç, and F.-J. Wu, “Group-in: Group inference from wireless traces of mobile devices,” in 2020 19th ACM/IEEE International Conference on Information Processing in Sensor Networks (IPSN). IEEE, 2020, pp. 157–168.
[22] S. He and K. G. Shin, “Crowd-flow graph construction and identification with spatio-temporal signal feature fusion,” in IEEE INFOCOM 2019-IEEE Conference on Computer Communications. IEEE, 2019, pp. 757–765.
[23] R. Sen, Y. Lee, K. Jayarajah, A. Misra, and R. K. Balan, “Grumon: Fast and accurate group monitoring for heterogeneous urban spaces,” in Proceedings of the 12th ACM conference on embedded network sensor systems, 2014, pp. 46–60.
[24] W.-z. Wang, Y.-w. Guo, B.-y. Huang, G.-r. Zhao, B.-q. Liu, and L. Wang, “Analysis of filtering methods for 3d acceleration signals in body sensor network,” in International Symposium on Bioelectronics and Bioinformations 2011. IEEE, 2011, pp. 263–266.
[25] C. Sadowski and G. Levin, “Simhash: Hash-based similarity detection,” Technical report, Google, 2007.
[26] “OpenStreetMap,” https://www.openstreetmap.org/copyright.