通信感知一体化系统中识别准确率和通信吞吐量的权衡分析

通信感知一体化（Integrated Sensing and Communication，ISAC）技术是指利用无线电磁波的随机性与确定性，实现通信与感知功能在时域、频域及空域上的复用。ISAC 系统的核心问题是通信-感知性能的可达域分析以及权衡分析。传统方法聚焦于模型驱动的感知任务，利用香农定理与估计理论推导其通-感性能可达域。但这类方法并不适用于数据驱动的感知任务。从技术角度看，数据驱动的感知任务更为复杂且需要语义特征的提取，因此高度依赖深度神经网络。这导致其性能难以通过数学建模分析手段获得，亟需通过数据驱动的方法来构建 ISAC 系统通信-感知性能的可达域，而这类方法在现有的 ISAC 研究中尚处于空白状态。

针对上述问题，本文以动作识别为例，提出了一个仿真驱动的性能预测器及优化器（Simulation-Driven Performance Predictor and oPtimizer，SDP3），以实现数据驱动的通信-感知性能权衡分析。SDP3 由 SDP3 数据仿真器、性能预测器和性能优化器组成。为了能够节省进行真实运动数据采集所消耗的大量人力与时间，本文在 SDP3 数据仿真器中提出了一个数据辅助的混合信道（Data-Assisted Hybrid Channel，DAHC）模型来模拟不同人体动作导致的信道响应。其中，本文使用了 Boulic 模型对人体的动作进行建模，从而实现对目标场景中的运动数据集的高效生成。SDP3 性能预测器利用深度时频图网络（Deep Spectrogram Network，DSN）对仿真器生成的包含运动信息的信道响应数据集进行训练和推理，从而得到相应的识别准确率；并利用多个具有闭式表达式的参数学习模型对识别准确率进行拟合，得到一个最优的模型来描述识别准确率与感知资源的关系。基于此，本文研究了一个优化通信与感知性能的资源分配问题，提出了对应的 SDP3 性能优化器，并给出了不同用户需求下的最优资源分配方案。进一步地，通过 SDP3 性能优化器，本文给出了识别准确率和通信吞吐量的可达域，包括一个通信饱和区、一个感知饱和区和通信-感知知对抗区。ISAC 系统理想的通信-感知平衡性能出现在第三个区域。此外，本文还论证了 SDP3 数据仿真器的保真性，即通过 DAHC 模型产生的信道响应数据集与实际实验获得的信道响应数据集在 KL 散度和识别准确率方面接近一致。最后，利用动作捕捉（Motion Capture，MoCap）技术生成的动作数据集进行训练与推理，并对得到的识别准确率进行了比较和分析。

Integrated Sensing and Communication (ISAC) is a technology that exploits the randomness and determinism of wireless electromagnetic waves to multiplex the communication and sensing functions in the time domain, frequency domain and space domain. The core issue of ISAC systems is the achievable region analysis and trade-off analysis of the communication-sensing performance. Traditional approaches focus on model-driven sensing tasks and derive their communication-sensing performance achievable regions by leveraging Shannon’s theorem and estimation theory. Nevertheless, such approaches are not applicable to data-driven sensing tasks. From a technical perspective, data-driven sensing tasks are more sophisticated and require semantic feature extraction, thus relying heavily on deep neural networks. This leads to the fact that their performances are difficult to be obtained by means of mathematical modeling analysis, and there is an urgent demand to construct achievable region for the communication-sensing performance of ISAC systems through data-driven methods, which is still a gap in the existing ISAC research.

To address the above issues, this paper adopts motion recognition as an analogy and proposes a Simulation-Driven Performance Predictor and oPtimizer (SDP3) to enable data-driven communication-sensing performance tradeoff analysis. SDP3 consists of the SDP3 data simulator, performance predictor, and performance optimizer. In order to save the labor and time consumed in real motion data acquisition, a Data-Assisted Hybrid Channel (DAHC) model is proposed in the SDP3 data simulator to simulate the channel response induced by different human motions. In particular, the Boulic model is adopted to model the human body’s movements, so as to achieve efficient generation of motion datasets in the target scenes. The SDP3 performance predictor leverages the Deep Spectrogram Network (DSN) to train and infer the channel response dataset containing motion information generated by the simulator to obtain the corresponding recognition accuracy; and fits the recognition accuracy with multiple parametric learning models featuring closed-form expressions to obtain an optimal model to describe the the relationship between recognition accuracy and sensory resources. On this basis, this paper investigates a resource allocation problem to optimize the sensing and communication performance, proposes the corresponding SDP3 performance optimizer, and provides the optimal resource allocation scheme for different user requirements. Furthermore, with the SDP3 performance optimizer, the achievable regions for recognition accuracy and communication throughput are given, including a communication saturation zone, a sensing saturation zone, and a communication-sensing adversarial zone. The ideal communication-sensing balance performance of the ISAC system appears in the third region. In addition, this paper demonstrates the fidelity of the SDP3 data simulator, where the channel response dataset generated by the DAHC model is in close agreement with the channel response dataset obtained from actual experiments in terms of KL Divergence and recognition accuracy. Finally, the motion datasets generated by Motion Capture (MoCap) technology is utilized for training and inference, and the obtained recognition accuracies are compared and analyzed.

[1] CUI Y, LIU F, JING X, et al. Integrating sensing and communications for ubiquitous IoT: Applications, trends, and challenges[J]. IEEE Network, 2021, 35(5): 158-167.
[2] ZHANG J A, LIU F, MASOUROS C, et al. An overview of signal processing techniques for joint communication and radar sensing[J]. IEEE Journal of Selected Topics in Signal Processing, 2021, 15(6): 1295-1315.
[3] LIU F, CUI Y, MASOUROS C, et al. Integrated sensing and communications: Towards dual-functional wireless networks for 6G and beyond[J]. IEEE Journal on Selected Areas in Communications, 2022, 40(6): 1728-1767.
[4] SHANNON C E. A mathematical theory of communication[J]. The Bell System Technical Journal, 1948, 27(3): 379-423.
[5] KAY S M. Fundamentals of statistical signal processing: estimation theory[M]. Prentice-Hall, Inc., 1993.
[6] TAN B, CHEN Q, CHETTY K, et al. Exploiting WiFi channel state information for residential healthcare informatics[J]. IEEE Communications Magazine, 2018, 56(5): 130-137.
[7] AMIN M G, ZHANG Y D, AHMAD F, et al. Radar signal processing for elderly fall detection: The future for in-home monitoring[J]. IEEE Signal Processing Magazine, 2016, 33(2): 71-80.
[8] MEALEY R M. A Method for Calculating Error Probabilities in a Radar Communication System[J/OL]. IEEE Transactions on Space Electronics and Telemetry, 1963, 9(2): 37-42. DOI: 10.1109/TSET.1963.4337601.
[9] REED J H, CLEGG A W, PADAKI A V, et al. On the Co-Existence of TD-LTE and Radar Over 3.5 GHz Band: An Experimental Study[J/OL]. IEEE Wireless Communications Letters, 2016, 5(4): 368-371. DOI: 10.1109/LWC.2016.2560179.
[10] RAO R M, MAROJEVIC V, REED J H. Semi-Blind Post-Equalizer SINR Estimation and Dual CSI Feedback for Radar-Cellular Coexistence[J/OL]. IEEE Transactions on Vehicular Technology, 2020, 69(9): 9720-9735. DOI: 10.1109/TVT.2020.3001911.
[11] CUI Y, KOIVUNEN V, JING X. Interference alignment based precoder-decoder design for radar-communication co-existence[C/OL]//2017 51st Asilomar Conference on Signals, Systems, and Computers. 2017: 1290-1295. DOI: 10.1109/ACSSC.2017.8335561.
[12] WANG X, FEI Z, ZHENG Z, et al. Joint Waveform Design and Passive Beamforming for RIS-Assisted Dual-Functional Radar-Communication System[J/OL]. IEEE Transactions on Vehicular Technology, 2021, 70(5): 5131-5136. DOI: 10.1109/TVT.2021.3075497.
[13] XU C, CLERCKX B, CHEN S, et al. Rate-Splitting Multiple Access for Multi-Antenna Joint Radar and Communications[J/OL]. IEEE Journal of Selected Topics in Signal Processing, 2021, 15(6): 1332-1347. DOI: 10.1109/JSTSP.2021.3110312.
[14] YUAN W, WEI Z, LI S, et al. Integrated Sensing and Communication-Assisted Orthogonal Time Frequency Space Transmission for Vehicular Networks[J/OL]. IEEE Journal of Selected Topics in Signal Processing, 2021, 15(6): 1515-1528. DOI: 10.1109/JSTSP.2021.3117404.
[15] LIU F, MASOUROS C, PETROPULU A P, et al. Joint Radar and Communication Design: Applications, State-of-the-Art, and the Road Ahead[J/OL]. IEEE Transactions on Communications, 2020, 68(6): 3834-3862. DOI: 10.1109/TCOMM.2020.2973976.
[16] XU C, CLERCKX B, ZHANG J. Multi-antenna joint radar and communications: Precoder optimization and weighted sum-rate vs probing power tradeoff[J]. IEEE Access, 2020, 8: 173974-173982.
[17] HAN L, WU K. Joint wireless communication and radar sensing systems–state of the art and future prospects[J]. IET Microwaves, Antennas & Propagation, 2013, 7(11): 876-885.
[18] LI X, ZHAO Q, GUAN X, et al. Sensing and Communication Tradeoff for Cognitive Access of Continues-Time Markov Channels[C/OL]//2010 IEEE Wireless Communication and Networking Conference. 2010: 1-6. DOI: 10.1109/WCNC.2010.5506649.
[19] HASSANIEN A, AMIN M G, ZHANG Y D, et al. Dual-Function Radar-Communications: Information Embedding Using Sidelobe Control and Waveform Diversity[J/OL]. IEEE Transactions on Signal Processing, 2016, 64(8): 2168-2181. DOI: 10.1109/TSP.2015.2505667.
[20] HASSANIEN A, AMIN M G, ZHANG Y D, et al. Non-coherent PSK-based dual-function radar-communication systems[C/OL]//2016 IEEE Radar Conference (RadarConf). 2016: 1-6. DOI: 10.1109/RADAR.2016.7485066.
[21] LIU F, MASOUROS C, LI A, et al. MU-MIMO Communications With MIMO Radar: From Co-Existence to Joint Transmission[J/OL]. IEEE Transactions on Wireless Communications, 2018, 17(4): 2755-2770. DOI: 10.1109/TWC.2018.2803045.
[22] KUMARI P, GONZALEZ-PRELCIC N, HEATH R W. Investigating the IEEE 802.11ad Standard for Millimeter Wave Automotive Radar[C/OL]//2015 IEEE 82nd Vehicular Technology Conference (VTC2015-Fall). 2015: 1-5. DOI: 10.1109/VTCFall.2015.7390996.
[23] GROSSI E, LOPS M, VENTURINO L, et al. Opportunistic Radar in IEEE 802.11ad Networks[J/OL]. IEEE Transactions on Signal Processing, 2018, 66(9): 2441-2454. DOI: 10.1109/TSP.2018.2813300.
[24] CHALISE B K, AMIN M G, HIMED B. Performance tradeoff in a unified passive radar and communications system[J]. IEEE Signal Processing Letters, 2017, 24(9): 1275-1279.
[25] CHALISE B K, HIMED B. Performance tradeoff in a unified multi-static passive radar and communication system[C]//2018 IEEE Radar Conference (RadarConf18). IEEE, 2018: 0653-0658.
[26] LIU F, LIU Y F, LI A, et al. Cramer-Rao Bound Optimization for Joint Radar-Communication Beamforming[J/OL]. IEEE Transactions on Signal Processing, 2021. DOI: 10.1109/TSP.2021.3135692.
[27] HUA H, SONG X, FANG Y, et al. MIMO integrated sensing and communication with extended target: CRB-rate tradeoff[C]//GLOBECOM 2022-2022 IEEE Global Communications Conference. IEEE, 2022: 4075-4080.
[28] CHIRIYATH A R, PAUL B, JACYNA G M, et al. Inner bounds on performance of radar and communications co-existence[J]. IEEE Transactions on Signal Processing, 2015, 64(2): 464-474.
[29] CHIRIYATH A R, PAUL B, BLISS D W. Radar-communications convergence: Coexistence, cooperation, and co-design[J]. IEEE Transactions on Cognitive Communications and Networking, 2017, 3(1): 1-12.
[30] LIU J, LIU H, CHEN Y, et al. Wireless sensing for human activity: A survey[J]. IEEE Communications Surveys & Tutorials, 2019, 22(3): 1629-1645.
[31] LIU X, HUANG T, SHLEZINGER N, et al. Joint transmit beamforming for multiuser MIMO communications and MIMO radar[J]. IEEE Transactions on Signal Processing, 2020, 68: 3929-3944.
[32] YANG Z, ZHOU Z, LIU Y. From RSSI to CSI: Indoor localization via channel response[J]. ACM Computing Surveys (CSUR), 2013, 46(2): 1-32.
[33] KOTARU M, JOSHI K, BHARADIA D, et al. Spotfi: Decimeter level localization using wifi[C]//Proceedings of the 2015 ACM Conference on Special Interest Group on Data Communication. 2015: 269-282.
[34] CHEN V C, LI F, HO S S, et al. Micro-Doppler effect in radar: phenomenon, model, and simulation study[J]. IEEE Transactions on Aerospace and Electronic Systems, 2006, 42(1): 2-21.
[35] KIM Y, LING H. Human activity classification based on micro-Doppler signatures using a support vector machine[J]. IEEE Transactions on Geoscience and Remote Sensing, 2009, 47(5): 1328-1337.
[36] KIM Y, MOON T. Human detection and activity classification based on micro-Doppler signatures using deep convolutional neural networks[J]. IEEE Geoscience and Remote Sensing Letters, 2015, 13(1): 8-12.
[37] ZHANG Y, ZHENG Y, QIAN K, et al. Widar3. 0: Zero-effort cross-domain gesture recognition with Wi-Fi[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 44(11): 8671-8688.
[38] ZENG Y, WU D, XIONG J, et al. FarSense: Pushing the range limit of WiFi-based respiration sensing with CSI ratio of two antennas[J]. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 2019, 3(3): 1-26.
[39] CHEN Q, CHETTY K, WOODBRIDGE K, et al. Signs of life detection using wireless passive radar[C]//2016 IEEE Radar Conference (RadarConf). IEEE, 2016: 1-5.
[40] ZHENG Y, YUE Z, CHENSHU W. AIoT 时代的智能无线感知：特征、算法、数据集[J].中国计算机学会通讯, 2020, 16(2): 50-56.
[41] JIANG W, XUE H, MIAO C, et al. Towards 3D human pose construction using WiFi[C]//Proceedings of the 26th Annual International Conference on Mobile Computing and Networking. 2020: 1-14.
[42] LI W, BOCUS M J, TANG C, et al. On CSI and Passive Wi-Fi Radar for Opportunistic Physical Activity Recognition[J/OL]. IEEE Transactions on Wireless Communications, 2022, 21(1): 607-620. DOI: 10.1109/TWC.2021.3098526.
[43] WANG F, GONG W, LIU J. On spatial diversity in WiFi-based human activity recognition: A deep learning-based approach[J]. IEEE Internet of Things Journal, 2018, 6(2): 2035-2047.
[44] WANG D, YANG J, CUI W, et al. Multimodal CSI-based human activity recognition using GANs[J]. IEEE Internet of Things Journal, 2021, 8(24): 17345-17355.
[45] 3GPP. Study on channel model for frequencies from 0.5 to 100 GHz[J]. 3GPP TR 38.901 V16.1.0, Dec. 2019.
[46] MALTSEV A, PUDEYEV A, LOMAYEV A, et al. Channel modeling in the next generation mmWave Wi-Fi: IEEE 802.11 ay standard[C]//European Wireless 2016; 22th European Wireless Conference. VDE, 2016: 1-8.
[47] KYOSTI P. WINNER II channel models[J]. IST, Tech. Rep. IST-4-027756 WINNER II D1. 1.2 V1. 2, 2007.
[48] NURMELA V, KARTTUNEN A, ROIVAINEN A, et al. Deliverable D1. 4 METIS channel models[J]. Proc. Mobile Wireless Commun. Enablers Inf. Soc.(METIS), 2015: 1.
[49] JAECKEL S, RASCHKOWSKI L, BÖRNER K, et al. QuaDRiGa: A 3-D multi-cell channel model with time evolution for enabling virtual field trials[J]. IEEE Transactions on Antennas and Propagation, 2014, 62(6): 3242-3256.
[50] JU S, KANHERE O, XING Y, et al. A millimeter-wave channel simulator NYUSIM with spatial consistency and human blockage[C]//2019 IEEE Global Communications Conference (GLOBECOM). IEEE, 2019: 1-6.
[51] BOULIC R, THALMANN N M, THALMANN D. A global human walking model with real-time kinematic personification[J]. The Visual Computer, 1990, 6(6): 344-358.
[52] Motion Research Laboratory, Carnegie Mellon University. Motion Capture Technology[EB/OL]. http://mocap.cs.cmu.edu.
[53] DOMHAN T, SPRINGENBERG J T, HUTTER F. Speeding up automatic hyperparameter optimization of deep neural networks by extrapolation of learning curves[C]//Twenty-fourth International Joint Conference on Artificial Intelligence. 2015.
[54] SALEH A A, VALENZUELA R. A statistical model for indoor multipath propagation[J]. IEEE Journal on Selected Areas in Communications, 1987, 5(2): 128-137.
[55] SEIDEL S Y, RAPPAPORT T S. Site-specific propagation prediction for wireless in-building personal communication system design[J]. IEEE Transactions on Vehicular Technology, 1994, 43(4): 879-891.
[56] VAN DORP P, GROEN F. Human walking estimation with radar[J]. IEE Proceedings-Radar, Sonar and Navigation, 2003, 150(5): 356-365.
[57] WANG Y, SHI Z, HUANG L, et al. An extension of spatial channel model with spatial consistency[C]//2016 IEEE 84th Vehicular Technology Conference (VTC-Fall). IEEE, 2016: 1-5.
[58] ORFANIDIS S J. Electromagnetic waves and antennas[M]. Rutgers University New Brunswick, NJ, 2002.
[59] CRISPIN J, MAFFETT A. Radar cross-section estimation for simple shapes[J]. Proceedings of the IEEE, 1965, 53(8): 833-848.
[60] ZHANG M, DU R, PENG X, et al. Channel models for WLAN sensing systems[R/OL]. 2021. https://mentor.ieee.org/802.11/documents?is_dcn=Meihong.
[61] OPPENHEIM A V. Discrete-time signal processing[M]. Pearson Education India, 1999.
[62] HE K, ZHANG X, REN S, et al. Identity mappings in deep residual networks[C]//Computer Vision–ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part IV 14. Springer, 2016: 630-645.
[63] WANG S, WU Y C, XIA M, et al. Machine intelligence at the edge with learning centric power allocation[J]. IEEE Transactions on Wireless Communications, 2020, 19(11): 7293-7308.
[64] ZHOU L, HONG Y, WANG S, et al. Learning centric wireless resource allocation for edge computing: Algorithm and experiment[J]. IEEE Transactions on Vehicular Technology, 2020, 70(1): 1035-1040.
[65] JOHNSON M, ANDERSON P, DRAS M, et al. Predicting accuracy on large datasets from smaller pilot data[C]//Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 2018: 450-455.
[66] LIU J, PORAT R, JINDAL N, et al. IEEE 802.11 ax channel model document[J]. Wireless LANs, Rep. IEEE, 2014: 802-11.