بهبود دقت تعيين موقعيت در مكان‌هاي سرپوشيده با استفاده از شبكه بي‌سيم با روش اثر مكاني محل

Improves the Accuracy of Indoors Positioning Using a Wireless Network

امروزه سيستم‌هاي تعيين موقعيت جهاني، در مواقعي كه ديد مستقيم بين كاربر و ماهواره‌هاي آن‌ها وجود ندارد، در درون ساختمان‌ها و در ناحيه‌هاي شهري متراكم كارايي چنداني ندارند. از اين‌ رو در سال‌هاي اخير سيستم تعيين موقعيت داخلي به‌طور قابل‌توجهي مورداستفاده قرارگرفته است. هدف اصلي اين تحقيق تعيين موقعيت در داخل ساختمان با شبكه بي‌سيم مي‌باشد. الگوريتم تعيين موقعيت داخلي شبكه بي‌سيم مي‌تواند به چند روش انجام شود كه به‌طورمعمول از روش اثر مكاني استفاده مي‌كنند. در اين تحقيق سيستم تعيين موقعيت داخلي برپايه دو روش نزديك‌ترين همسايه و شبكه عصبي مصنوعي در سه سناريو پياده‌سازي شده است. به‌منظور افزايش دقت مكاني و كاهش زمان محاسبه، انتخاب فرستنده‌هاي مناسب و افزايش نقاط مرجع مورد توجه قرار گرفت. در روش نزديك‌ترين همسايگي با افزايش نقاط به ميزان دو برابر با روش درون‌يابي ميانگين، ميزان دقت هم به ميزان تقريباً دو برابر بهبود يافته است كه در محيط‌هاي مختلف با تعداد مختلف نقاط مرجع و پراكندگي مختلف اين نتيجه حاصل شده است. در محيط‌هايي همانند سناريو 3 كه نقاط مرجع به تعداد بالاتر و با پراكندگي يكسان‌تري پخش شده‌اند دقت تعيين موقعيت در روش نزديك‌ترين همسايگي بهبود يافته بهتر از شبكه عصبي بهبود يافته است و اين دقت براي نزديك‌ترين همسايگي در 85 درصد موارد زير 1 متر بوده كه در شبكه عصبي در 75 درصد موارد زير 1 متر مي‌باشد؛ اما در محيط‌هاي كه نقاط مرجع داراي تعداد كمتر مي‌باشد و پراكندگي يكساني ندارد روش شبكه عصبي بهبود يافته نتايج بهتري را نسبت به روش نزديك‌ترين همسايگي خواهد داشت. به‌طوري كه دقت نزديك‌ترين همسايگي بهبود يافته در سناريو 2 در 72 درصد موارد زير 2 متر بوده كه در شبكه عصبي بهبود يافته در 84 درصد موارد زير دقت 2 متر مي‌باشد.

Indoors positioning requires methods that can accurately represent the user's position. Unfortunately, Global Positioning System (GPS) due to signal attenuation when there are not the direct lines of sight from a mobile phone to least three satellites, especially in dense urban areas and inside the building cannot be used effectively. Internal positioning is based on four methods: time of arrival (TOA), time difference of arrival signal (DTOA), signal angle of arrival (AOA) and signal strength (RSS). Time of arrival, sometimes called time of flight (ToF), is the travel time of a radio signal from a single transmitter to a remote single receiver. This method is used the absolute time of arrival at a certain base station rather than the measured time difference between departing from one and arriving at the other station. The distance can be directly calculated from the time of arrival as signals travel with a known velocity. Time of arrival data from two base stations will narrow a position to a position circle; data from a third base station is required to resolve the precise position to a single point. Many radiolocation systems, including GPS, use ToA. The time difference of arrival (TDOA) is based on trilateration and uses the time difference measured from two stations to define a hyperbolic curve as an estimation of the position of the mobile device. Using an extra station provides a new hyperbolic and the intersection of two hyperbolic curves give the position of the mobile device. The angle of arrival (AOA) method is based on triangulation and determines the position of a mobile device using intersection of directional lines between the mobile device and at least two base stations. The RSS (Received Signal Strength) sometimes referred as RSSI (Received signal strength indicator) is a measurement of the power present in a received radio signal. The nodes used by the accurate Wi-Fi location monitor and Bluetooth beacon Tracker are capable of measuring the RSS of nearby Wi-Fi and BLE devices. One of the methods of determining the position based on the RSS method is the wireless network signal, which is now used extensively in a local area network. An internal positioning algorithm for a wireless network can be done in three ways: proximity algorithm, triangulation algorithms and sense analysis algorithms. Due to the multi-path effects and the effects of the signal propagation inside the internal environments and the lack of direct vision between the transmitter and the receiver, sense analysis algorithm is used, which is usually based on fingerprint location using signal strength. Fingerprint refers to the way adaptation some of the characteristics of the signal that is dependent on the location. There are at least 3 positioning algorithms based on the fingerprint that use the sense analysis method: probabilistic, nearest neighbor, neural networks. In this paper, the positioning system is implemented based on all three methods and their accuracy is compared with each other. In each of these methods, in order to improve the accuracy of location and time of calculation, the choice of suitable transmitters and the increase of reference points by different methods are compared and their accuracy is compared with the methods of determining position in the usual mode.