{"id":2,"date":"2024-03-20T13:04:51","date_gmt":"2024-03-20T13:04:51","guid":{"rendered":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/?page_id=2"},"modified":"2025-04-01T13:46:03","modified_gmt":"2025-04-01T04:46:03","slug":"contents","status":"publish","type":"page","link":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/contents\/","title":{"rendered":"\u7814\u7a76\u5185\u5bb9 (Research Contents)"},"content":{"rendered":"<h4>\u7d71\u8a08\u7684\u591a\u6b21\u5143\u4fe1\u53f7\u51e6\u7406\u624b\u6cd5\u306b\u57fa\u3065\u304f\u4fe1\u53f7\u691c\u51fa<\/h4>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-medium wp-image-221\" src=\"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/wp-content\/uploads\/2024\/04\/mimo-300x257.png\" alt=\"\" width=\"300\" height=\"257\" srcset=\"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/wp-content\/uploads\/2024\/04\/mimo-300x257.png 300w, http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/wp-content\/uploads\/2024\/04\/mimo-333x286.png 333w, http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/wp-content\/uploads\/2024\/04\/mimo.png 599w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/><\/p>\n<h5>\u3000\u5927\u898f\u6a21MIMO (Multi-Input Multi-Output)\u200b<\/h5>\n<h6>\u3000Massive MIMO<\/h6>\n<p>\u5927\u898f\u6a21MIMO\u306f\uff0c\u9001\u53d7\u4fe1\u6a5f\u3067\u591a\u6570\u306e\u30a2\u30f3\u30c6\u30ca\u3092\u7528\u3044\u3066\u60c5\u5831\u3092\u7a7a\u9593\u7684\u306b\u591a\u91cd\u3057\uff0c\u5927\u91cf\u306e\u60c5\u5831\u3092\u4f1d\u9001\u3059\u308b5G\u306e\u4e2d\u6838\u6280\u8853\u3067\u3059\uff0e\u3055\u3089\u306bIoT\u30c7\u30d0\u30a4\u30b9\u306e\u666e\u53ca\u3092\u80cc\u666f\u306b\uff0c\u5927\u91cf\u306e\u7aef\u672b\u304b\u3089\u306e\u30c7\u30fc\u30bf\u96c6\u7d04\u3092\u76ee\u7684\u3068\u3057\u305f\u5927\u898f\u6a21\u30de\u30eb\u30c1\u30e6\u30fc\u30b6MIMO \u306e\u9700\u8981\u304c\u9ad8\u307e\u3063\u3066\u3044\u307e\u3059.<\/p>\n<p>\u3057\u304b\u3057\u3053\u306e\u5834\u5408\uff0c\u591a\u6570\u306e\u30e6\u30fc\u30b6\u60c5\u5831\u3092\u53d7\u4fe1\u6a5f\u3067\u540c\u6642\u306b\u51e6\u7406\u3059\u308b\u5fc5\u8981\u304c\u3042\u308b\u305f\u3081\uff0c\u5f93\u6765\u306e\u4fe1\u53f7\u51e6\u7406\u624b\u6cd5\u3067\u306f\u5b9f\u73fe\u4e0d\u53ef\u80fd\u306a\u51e6\u7406\u6f14\u7b97\u91cf\u3068\u306a\u3063\u3066\u3057\u307e\u3046\u554f\u984c\u304c\u3042\u308a\u307e\u3059\uff0e<\/p>\n<p>Massive MIMO is a promising technology for Beyond 5G and 6G, which uses a large number of antennas at the transmitter and receiver to spatially multiplex information and transmit large amounts of information. Furthermore, with the spread of IoT devices, demands for large-scale multi-user MIMO become higher to gather data from plenty of terminals.<\/p>\n<p>However, the massive MIMO channel requires the receiver to separate large amounts of data from multiple users simultaneously, which leads to an increase in the calculational cost that the conventional signal processing method cannot handle.<\/p>\n<h5>\u3000\u201cMore is different.\u201d (\u91cf\u304c\u5897\u3048\u308b\u3068\u8cea\u304c\u5909\u308f\u308b)\u200b<\/h5>\n<p>\u6271\u3046\u4fe1\u53f7\u306e\u591a\u6b21\u5143\u5316\u306f\uff0c\u672c\u6765\uff0c\u4fe1\u53f7\u51e6\u7406\u91cf\u306e\u5897\u52a0\u3092\u62db\u304f\u3082\u306e\u3067\u3059\uff0e\u3057\u304b\u3057\u305d\u306e\u6b21\u5143\u6570\u304c\u975e\u5e38\u306b\u5927\u304d\u304f\u306a\u308b\u3068\uff0c\u9762\u767d\u3044\u3053\u3068\u306b\uff0c\u51e6\u7406\u91cf\u3092\u5927\u304d\u304f\u6e1b\u3089\u3057\u3064\u3064\u9ad8\u3044\u6027\u80fd\u3092\u9054\u6210\u3059\u308b\u4fe1\u53f7\u5206\u96e2\u624b\u6cd5\u3092\u69cb\u6210\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\uff0e\u200b<\/p>\n<p>\u3053\u308c\u3089\u306e\u624b\u6cd5\u306f\u4fe1\u5ff5\uff08\u78ba\u7387\uff09\u4f1d\u642c\u6cd5 (BP; Belief Propagation) \u3068\u3044\u3046\u30af\u30e9\u30b9\u306b\u5c5e\u3059\u308b\u4e00\u9023\u306e\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3092\u4fe1\u53f7\u51e6\u7406\u5206\u91ce\u306b\u5fdc\u7528\u3057\u305f\u3082\u306e\u3067\u3042\u308a\uff0c\u591a\u6570\u306e\u8cea\u70b9\u306e\u96c6\u56e3\u7684\u306a\u6319\u52d5\u3092\u4e0e\u3048\u308b\u71b1\u529b\u5b66\u30fb\u7d71\u8a08\u529b\u5b66\u306e\u77e5\u898b\u3092\u57fa\u790e\u3068\u3057\u3066\u3044\u307e\u3059. BP\u306b\u57fa\u3065\u304f\u4fe1\u53f7\u51e6\u7406\u306b\u306f\u30c7\u30fc\u30bf\u89e3\u6790\u3092\u76ee\u7684\u3068\u3057\u305f\u4eba\u5de5\u77e5\u80fd\u5206\u91ce\u3067\u63d0\u6848\u3055\u308c\u305f\u624b\u6cd5\u3068\u306e\u985e\u4f3c\u6027\u3082\u898b\u51fa\u3055\u308c\u3066\u3044\u307e\u3059\uff0e\u3053\u306e\u3088\u3046\u306b\uff0c\u9ad8\u5ea6\u306a\u7121\u7dda\u4fe1\u53f7\u51e6\u7406\u306e\u5b9f\u73fe\u306b\u306f\uff0c\u5206\u91ce\u6a2a\u65ad\u7684\u306a\u8996\u70b9\u304c\u5fc5\u8981\u4e0d\u53ef\u6b20\u3067\u3059\uff0e\u672c\u7814\u7a76\u306e\u76ee\u7684\u306f\uff0c\u3088\u308a\u52b9\u7387\u7684\u306a\u591a\u6b21\u5143\u4fe1\u53f7\u51e6\u7406\u306e\u5275\u51fa\u3067\u3042\u308a\uff0c\u305d\u306e\u6210\u679c\u306f\u672c\u8cea\u7684\u306b\u540c\u3058\u554f\u984c\u3092\u6271\u3046\u5e45\u5e83\u3044\u5206\u91ce\u3078\u306e\u8ca2\u732e\u3068\u306a\u308a\u307e\u3059\uff0e\u200b<\/p>\n<p>The increase in the dimensionality of the signals is inherently disadvantageous in terms of signal processing. However, as the number of dimensions tends to be very large, such multidimensionality can be conversely exploited to construct signal separation methods to achieve high performance with significantly reduced computational complexity.<\/p>\n<p>Such methods stem from a series of algorithms belonging to the class of Belief Propagation (BP) and work based on the theoretical knowledge of thermodynamics and statistical mechanics. In addition, these methods show some similarities with the ones proposed in the field of machine learning for data analysis. That means our research needs a cross-cutting perspective, and our achievements in the context of wireless communication contribute to a broader range of fields dealing with essentially the same problem.<\/p>\n<h5>\u3000\u6a5f\u68b0\u5b66\u7fd2\u3068\u306e\u878d\u5408 (\u7406\u60f3\u3068\u73fe\u5b9f\u306e\u305a\u308c\u306e\u30e2\u30c7\u30eb\u5316)\u200b<\/h5>\n<h6>\u3000Fusion of statistical multidimensional signal processing and machine learning (Modeling the error occurring in the practical scenario)<\/h6>\n<p>\u3053\u3046\u3057\u305f\u591a\u6b21\u5143\u4fe1\u53f7\u51e6\u7406\u3092\u7121\u7dda\u901a\u4fe1\u30b7\u30b9\u30c6\u30e0\u3067\u5229\u7528\u3059\u308b\u306b\u306f\uff0c\u4e57\u308a\u8d8a\u3048\u308b\u3079\u304d\u8ab2\u984c\u304c\u3042\u308a\u307e\u3059\uff0e\u305d\u308c\u306f\uff0c\u73fe\u5b9f\u306e\u7121\u7dda\u74b0\u5883\u304c\uff0c\u7d71\u8a08\u529b\u5b66\u3084\u30d3\u30c3\u30b0\u30c7\u30fc\u30bf\u89e3\u6790\u3067\u4e0e\u3048\u3089\u308c\u308b\u3088\u3046\u306a\u7406\u60f3\u7684\u306a\u5927\u898f\u6a21\u30e2\u30c7\u30eb\u306b\u306a\u3063\u3066\u3044\u306a\u3044\u3053\u3068\u3067\u3059\uff0e\u9001\u53d7\u4fe1\u30a2\u30f3\u30c6\u30ca\u6570\u304c\u7cbe\u3005\u6570\u5341\u304b\u3089\u767e\u3068\u3044\u3046\u306e\u306f\uff0c\u524d\u8ff0\u306e\u5206\u91ce\u3067\u3044\u3046\u3068\u5c11\u306a\u3059\u304e\u307e\u3059\u3057\uff0c\u53d7\u4fe1\u4fe1\u53f7\u306e\u89b3\u6e2c\u76f8\u95a2\u6027\u3084\u91cf\u5b50\u5316\u8aa4\u5dee\u306a\u3069\uff0c\u7121\u7dda\u7279\u6709\u306e\u554f\u984c\u3082\u3042\u308a\u307e\u3059\uff0e\u3053\u306e\u3088\u3046\u306a\u30e2\u30c7\u30eb\u8aa4\u5dee\u306f\u6570\u7406\u7684\u306a\u30e2\u30c7\u30eb\u5316\u304c\u96e3\u3057\u304f\uff0c\u3053\u308c\u307e\u3067\u6709\u52b9\u306a\u624b\u7acb\u3066\u306f\u591a\u304f\u306f\u3042\u308a\u307e\u305b\u3093\u3067\u3057\u305f\uff0e<\/p>\n<p>\u3057\u304b\u3057\uff0c\u8fd1\u5e74\u306e\u8a08\u7b97\u6a5f\u6027\u80fd\u306e\u5411\u4e0a\u306b\u4f34\u3046\u6df1\u5c64\u5b66\u7fd2\u6280\u8853\u306e\u767a\u5c55\u306b\u4f34\u3044\uff0c\u3053\u306e\u9694\u305f\u308a\u3092\u57cb\u3081\u308b\u89e3\u6c7a\u306e\u7cf8\u53e3\u304c\u898b\u51fa\u3055\u308c\u3064\u3064\u3042\u308a\u307e\u3059\uff0e\u7d71\u8a08\u7684\u591a\u6b21\u5143\u4fe1\u53f7\u51e6\u7406\u3068\u6a5f\u68b0\u5b66\u7fd2\u306e\u878d\u5408\u306f\u65b0\u3057\u3044\u7814\u7a76\u5206\u91ce\u3067\u3042\u308a\uff0c\u73fe\u5728\u306e\u30db\u30c3\u30c8\u30c8\u30d4\u30c3\u30af\u3067\u3059\uff0e<\/p>\n<p>The practical use of such multidimensional signal processing in wireless communication systems poses a challenge to overcome. Unlike statistical mechanics and big data analysis, the practical wireless channel cannot be regarded as an ideal model. Compared to these fields, the number of dimensions in the wireless channel (i.e., the number of transmitting and receiving antennas, which is a few dozen to a hundred at best) is too small. There are also several problems unique to wireless communication, such as spatial correlation and quantization errors in the received signal. Such model errors are difficult to describe via mathematical models, and so far, very few effective ways to deal with them have been suggested.<\/p>\n<p>However, thanks to the recent advances in deep learning techniques, solutions are being found to bridge this gap. The fusion of statistical multidimensional signal processing and machine learning is a frontier of our research and a current hot topic.<\/p>\n<h5>\u3000\u30b7\u30b9\u30c6\u30e0\u30ec\u30d9\u30eb\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3 (\u7406\u8ad6\u3068\u5fdc\u7528\u306e\u6a4b\u6e21\u3057)<\/h5>\n<h6>\u3000System-level Simulation (Bridging Theory and Application)<\/h6>\n<p>\u307e\u305f\uff0c\u524d\u8ff0\u306e\u4fe1\u53f7\u51e6\u7406\u624b\u6cd5\u306f\uff0c\u4fe1\u53f7\u691c\u51fa\u5668\u306e\u7d14\u7c8b\u306a\u6027\u80fd\u3092\u8a55\u4fa1\u3059\u308b\u305f\u3081\u306b\u884c\u308f\u308c\u308b\u30ea\u30f3\u30af\u30ec\u30d9\u30eb\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u306b\u304a\u3044\u3066\u306f\u6f14\u7b97\u91cf\u3068\u691c\u51fa\u7cbe\u5ea6\u306e\u826f\u597d\u306a\u30c8\u30ec\u30fc\u30c9\u30aa\u30d5\u3092\u9054\u6210\u3059\u308b\u3053\u3068\u304c\uff0c\u69d8\u3005\u306a\u7406\u8ad6\u7814\u7a76\u3092\u901a\u3057\u3066\u660e\u3089\u304b\u306b\u306a\u3063\u3066\u3044\u307e\u3059. \u3057\u304b\u3057\uff0c\u30ea\u30f3\u30af\u30ec\u30d9\u30eb\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u3067\u306f\u5b9f\u969b\u306e\u7121\u7dda\u901a\u4fe1\u74b0\u5883\u306b\u5bfe\u3057\u3066\u69d8\u3005\u306a\u6761\u4ef6\u304c\u7406\u60f3\u5316\u3055\u308c\u305f\u72b6\u6cc1\u3067\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\u304c\u884c\u308f\u308c\u3066\u304a\u308a\uff0c\u3053\u308c\u3089\u306e\u4fe1\u53f7\u51e6\u7406\u6280\u8853\u306e\u767a\u5c55\u304c\u5b9f\u969b\u306e\u7121\u7dda\u901a\u4fe1\u30b7\u30b9\u30c6\u30e0\u5168\u4f53\u306e\u6027\u80fd\u306b\u5bfe\u3057\u3066\u3069\u308c\u3060\u3051\u5bc4\u4e0e\u3067\u304d\u308b\u304b\u306b\u95a2\u3057\u3066\u306f\uff0c\u5225\u9014\u8a55\u4fa1\u304c\u5fc5\u8981\u3068\u306a\u308a\u307e\u3059.\u3000\u3068\u3053\u308d\u304c\uff0c\u3053\u3046\u3057\u305f\u5b9f\u7528\u5316\u306e\u89b3\u70b9\u306b\u57fa\u3065\u304f\u7d71\u8a08\u7684\u591a\u6b21\u5143\u4fe1\u53f7\u51e6\u7406\u624b\u6cd5\u306e\u8a55\u4fa1\u306f\uff0c\u5341\u5206\u306b\u884c\u308f\u308c\u3066\u3044\u306a\u3044\u306e\u304c\u73fe\u72b6\u3067\u3059.<\/p>\n<p>\u305d\u3053\u3067\u672c\u7814\u7a76\u5ba4\u3067\u306f\uff0c\u524d\u8ff0\u306e\u4fe1\u53f7\u691c\u51fa\u5668\u306e\u4f7f\u7528\u3092\u60f3\u5b9a\u3057\u305f\u30b7\u30b9\u30c6\u30e0\u30ec\u30d9\u30eb\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\uff08\u5b9f\u969b\u306e\u7121\u7dda\u901a\u4fe1\u30b7\u30b9\u30c6\u30e0\u3092\u6a21\u3057\u305f\u30b7\u30df\u30e5\u30ec\u30fc\u30b7\u30e7\u30f3\uff09\u3092\u69cb\u7bc9\u3059\u308b\u3053\u3068\u306b\u3088\u308a\uff0c\u7d71\u8a08\u7684\u4fe1\u53f7\u51e6\u7406\u306e\u5f37\u307f\u3092\u5341\u5206\u306b\u767a\u63ee\u3059\u308b\u305f\u3081\u306e\u7121\u7dda\u901a\u4fe1\u30b7\u30b9\u30c6\u30e0\u306e\u8a2d\u8a08\u898f\u7bc4\u3092\u63d0\u6848\u3059\u308b\u305f\u3081\u306e\u7814\u7a76\u3092\u884c\u3063\u3066\u304a\u308a\u307e\u3059.<\/p>\n<p>Various theoretical studies have shown that these signal-processing methods can achieve a good trade-off between computational complexity and detection precision in link-level simulations. However, in link-level simulations, some of the conditions are idealized to assess the pure performance of signal detectors, and it is another issue to find out to what extent these signal detectors can help improve the whole wireless communication system, yet such performance evaluations have not been made sufficiently.<\/p>\n<p>We try to design a system-level simulation (a simulation that mimics the whole wireless communication system) involving the statistical signal detectors to propose novel design norms for wireless communication systems that can fully exploit its advantages.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u7d71\u8a08\u7684\u591a\u6b21\u5143\u4fe1\u53f7\u51e6\u7406\u624b\u6cd5\u306b\u57fa\u3065\u304f\u4fe1\u53f7\u691c\u51fa \u3000\u5927\u898f\u6a21MIMO (Multi-Input Multi-Output)\u200b \u3000Massive MIMO \u5927\u898f\u6a21MIMO\u306f\uff0c\u9001\u53d7\u4fe1\u6a5f\u3067\u591a\u6570\u306e\u30a2\u30f3\u30c6\u30ca\u3092\u7528\u3044\u3066\u60c5\u5831\u3092\u7a7a\u9593\u7684\u306b\u591a\u91cd\u3057\uff0c\u5927\u91cf\u306e\u60c5\u5831\u3092\u4f1d\u9001\u3059\u308b5G\u306e\u4e2d\u6838\u6280\u8853\u3067\u3059\uff0e\u3055\u3089\u306bIoT\u30c7\u30d0\u30a4\u30b9\u306e\u666e\u53ca\u3092\u80cc\u666f\u306b\uff0c\u5927\u91cf\u306e\u7aef\u672b\u304b\u3089\u306e\u30c7\u30fc\u30bf\u96c6\u7d04\u3092\u76ee\u7684\u3068\u3057\u305f\u5927\u898f\u6a21\u30de\u30eb\u30c1\u30e6\u30fc\u30b6MIMO \u306e\u9700\u8981\u304c\u9ad8\u307e\u3063\u3066\u3044\u307e\u3059. \u3057\u304b\u3057\u3053\u306e\u5834\u5408\uff0c\u591a\u6570\u306e\u30e6\u30fc\u30b6\u60c5\u5831\u3092\u53d7\u4fe1\u6a5f\u3067\u540c\u6642\u306b\u51e6\u7406\u3059\u308b\u5fc5\u8981\u304c\u3042\u308b\u305f\u3081\uff0c\u5f93\u6765\u306e\u4fe1\u53f7\u51e6\u7406\u624b\u6cd5\u3067\u306f\u5b9f\u73fe\u4e0d\u53ef\u80fd\u306a\u51e6\u7406\u6f14\u7b97\u91cf\u3068\u306a\u3063\u3066\u3057\u307e\u3046\u554f\u984c\u304c\u3042\u308a\u307e\u3059\uff0e Massive MIMO is a promising technology for Beyond 5G and 6G, which uses a large number &hellip;<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"_links":{"self":[{"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/pages\/2"}],"collection":[{"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/comments?post=2"}],"version-history":[{"count":25,"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/pages\/2\/revisions"}],"predecessor-version":[{"id":575,"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/pages\/2\/revisions\/575"}],"wp:attachment":[{"href":"http:\/\/www2a.comm.eng.osaka-u.ac.jp\/ochiailab\/index.php\/wp-json\/wp\/v2\/media?parent=2"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}