{"id":5865,"date":"2014-09-24T15:40:54","date_gmt":"2014-09-24T22:40:54","guid":{"rendered":"http:\/\/palblog.fxpal.com\/?p=5865"},"modified":"2014-09-24T12:38:09","modified_gmt":"2014-09-24T19:38:09","slug":"loco-a-framework-for-indoor-location-of-mobile-devices","status":"publish","type":"post","link":"https:\/\/blog.fxpal.net\/?p=5865","title":{"rendered":"LoCo: a framework for indoor location of mobile devices"},"content":{"rendered":"<p>Last year, we initiated the LoCo project on indoor location. \u00a0The <a href=\"http:\/\/www.fxpal.com\/research-projects\/loco\/\">LoCo page<\/a> has more information, but our central goal is to provide highly accurate, room-level location information to enable indoor location services to complement the location services built on GPS outdoors.<\/p>\n<p>Last week, we presented our initial\u00a0<a href=\"http:\/\/www.fxpal.com\/publications\/loco_a_ready-to-deploy_framework_for_ef%EF%AC%81cient_room_localization_using_wi-fi\/\">results<\/a> on the work at <a href=\"http:\/\/ubicomp.org\/ubicomp2014\/\">Ubicomp 2014<\/a>. \u00a0In our\u00a0paper, we introduce a new approach to room-level location based on supervised classification. \u00a0Specifically, we use <a href=\"https:\/\/en.wikipedia.org\/wiki\/Boosting_(machine_learning)\">boosting<\/a>\u00a0in a one-versus-all formulation to enable highly accurate classification based on simple features derived from Wi-Fi received signal strength (RSSI) measures. \u00a0This approach offloads the bulk of the complexity to an offline training procedure, and the resulting classifier is sufficiently simple to be run on a mobile client directly. \u00a0We use a simple and robust feature set based on pairwise RSSI margin to both address Wi-Fi RSSI volatility.<\/p>\n<p style=\"font-size: 27px;\" align=\"center\"><img decoding=\"async\" src=\"https:\/\/s0.wp.com\/latex.php?latex=h_m%28X%29+%3D+%5Cbegin%7Bcases%7D+1+%26+X%28b_m%5E%7B%281%29%7D%29+-+X%28b_m%5E%7B%282%29%7D%29+%5Cgeq+%5Ctheta_m+%5C%5C+0+%26+%5Ctext%7Botherwise%7D+%5Cend%7Bcases%7D&#038;bg=ffffff&#038;fg=000&#038;s=0&#038;c=20201002\" alt=\"h_m(X) = &#92;begin{cases} 1 &amp; X(b_m^{(1)}) - X(b_m^{(2)}) &#92;geq &#92;theta_m &#92;&#92; 0 &amp; &#92;text{otherwise} &#92;end{cases}\" class=\"latex\" \/><\/p>\n<p style=\"text-align: left;\">The equation above shows an example <em>weak<\/em> learner which simply looks at two elements in an RSSI scan and compares their difference against a threshold. \u00a0The final <em>strong<\/em> classifier for each room is a weighted combination of a set of weak learners greedily selected to discriminate that room. \u00a0The feature is designed to express the ordering of RSSI values observed for specific access points, and a flexible reliance on the difference between them, and the threshold <img decoding=\"async\" src=\"https:\/\/s0.wp.com\/latex.php?latex=%5Ctheta_m&#038;bg=ffffff&#038;fg=000&#038;s=0&#038;c=20201002\" alt=\"&#92;theta_m\" class=\"latex\" \/> is determined in training. \u00a0An additional benefit of this choice is that processing a subset of the RSSI scan according to the selected weak learners further reduces the required computation. \u00a0Comparing against the kNN matching approach used in RedPin <a href=\"http:\/\/dl.acm.org\/citation.cfm?id=1410025\">[Bolliger, 2008]<\/a>, our results show competitive performance with substantially reduced complexity. \u00a0The Table below shows cross validation results from the paper for two data sets collected in\u00a0our office. \u00a0The classification time appears in the rightmost column.<\/p>\n<p><a href=\"http:\/\/palblog.fxpal.com\/wp-content\/uploads\/2014\/09\/ubicomp_table1.png\"><img decoding=\"async\" loading=\"lazy\" width=\"1477\" height=\"282\" class=\"aligncenter wp-image-5878\" src=\"http:\/\/palblog.fxpal.com\/wp-content\/uploads\/2014\/09\/ubicomp_table1.png\" alt=\"\" srcset=\"https:\/\/blog.fxpal.net\/wp-content\/uploads\/2014\/09\/ubicomp_table1.png 1477w, https:\/\/blog.fxpal.net\/wp-content\/uploads\/2014\/09\/ubicomp_table1-300x57.png 300w, https:\/\/blog.fxpal.net\/wp-content\/uploads\/2014\/09\/ubicomp_table1-1024x195.png 1024w\" sizes=\"(max-width: 1477px) 100vw, 1477px\" \/><\/a><a href=\"http:\/\/palblog.fxpal.com\/wp-content\/uploads\/2014\/09\/ubicomp_table.png\"><br \/>\n<\/a><\/p>\n<p>We are excited about the early progress we&#8217;ve made on this project and look forward to\u00a0building out our indoor location system in several directions in the near future. \u00a0But more than that, we look forward to building new location driven applications exploiting\u00a0this technique which can leverage existing infrastructure (Wi-Fi networks) and devices (cell phones) we already use.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Last year, we initiated the LoCo project on indoor location. \u00a0The LoCo page has more information, but our central goal is to provide highly accurate, room-level location information to enable indoor location services to complement the location services built on GPS outdoors. Last week, we presented our initial\u00a0results on the work at Ubicomp 2014. \u00a0In [&hellip;]<\/p>\n","protected":false},"author":50,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[7,39],"tags":[],"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=\/wp\/v2\/posts\/5865"}],"collection":[{"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=\/wp\/v2\/users\/50"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=5865"}],"version-history":[{"count":13,"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=\/wp\/v2\/posts\/5865\/revisions"}],"predecessor-version":[{"id":5869,"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=\/wp\/v2\/posts\/5865\/revisions\/5869"}],"wp:attachment":[{"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=5865"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=5865"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.fxpal.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=5865"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}