Introduction to the
Today, I will introduce the principles and implementation methods of indoor positioning.
WIFI Fidelity, also known as “mobile hotspot” in Chinese, is a trademark of the manufacturer of Wi-Fi Alliance as a brand certification of products. It is a wireless LAN technology created in the IEEE 802.11 standard. Due to the close correlation between the two systems, Wi-Fi is often used as a synonym for the IEEE 802.11 standard. “Wi-fi” is often written as “WiFi” or “WiFi,” but these are not recognized by the Wi-Fi Alliance. Its biggest advantage is the high transmission speed, up to 11Mbps, in addition to its effective distance is also very long, and it is compatible with various existing 802.11DSSS devices.
The essence of wifi positioning is actually base station positioning, which relies on wifi AP to conduct positioning by radiating gradually weakening fingerprint with itself as the center.
Wifi location background and significance
GPS is difficult to solve some positioning problems in indoor environments. WiFi exists in most indoor environments. Therefore, it is a very cost-saving method to use WiFi for positioning without additional hardware deployment. However, WiFi is not specifically designed for positioning, and traditional positioning methods based on time and Angle are not applicable to WiFi. In the past decade, location fingerprinting has been widely studied and adopted in indoor WiFi positioning. This paper reviews WiFi location fingerprinting, presents the challenges in this field, introduces the latest research, and provides some practical guidance.
The basic principle of
What is wifi fingerprint
Location fingerprinting associates locations in the physical environment with a kind of “fingerprint,” and each location corresponds to a unique fingerprint. The fingerprint can be one-dimensional or multidimensional, for example, when the device to be located is receiving or sending a message, then the fingerprint can be one or more features of the message or signal (the most common is signal strength). If the device is sending a signal, some fixed receiving device senses the signal or information of the device and locates it, which is often called remote location or network location. If a device to be located receives some signals or information from a fixed transmitting device, and then estimates its position according to these detected features, this method can be called self-positioning. The mobile device to be located may communicate the features it detects to a server node in the network, which can use all the information it can obtain to estimate the location of the mobile device. , this method can be called hybrid positioning. In all of these ways, the perceived signal characteristics need to be matched to the signal characteristics in a database, and the process can be regarded as a pattern recognition problem.
What does wifi fingerprint consist of
Location fingerprints can be of many types, and any “location unique” feature can be used as a location fingerprint. Signal multipath structure such as a position, a position on whether can detect the access point or base station, a location on the detected signals from the base station of RSS (received signal strength), a location on the round-trip time or delay of communication signal can position as a fingerprint, or it can be combined as a fingerprint.
Wifi fingerprint positioning method
Wifi positioning methods can be basically divided into two categories:
Is not based on RSSI
- TOA (time ofarrival)
- TDOA (Time Difference of Arrival)
- AOA (Angle of Arrival)
Based on RSSI
On smart phones, RSSI of signal strength and AP address sent by each Access Point (AP) around can be obtained through the system SDK. It seems to be the most feasible method to use RSSI for positioning at present. Therefore, the following focuses on two algorithms based on RSSI positioning: triangulation algorithm and fingerprint algorithm.
Localization algorithm
- Random method
Set the predicted position of the Test Point (TP) to that of any Reference Point (RP).
- KNN method
The main idea of KNN method is to find the K RPS with the highest WiFi fingerprint similarity with TP, and then average the corresponding positions of these K RPS, and finally get the position of TP.
- Prob method
The main idea of Prob method is to calculate the posterior probability of TP appearing on each RP through Bayesian formula, take K RP with the highest probability, average the corresponding positions of these K RP, and finally get the position of TP.
- Stg method
Stg method is an improved method of KNN method. It filters RP through K signal access points of TP with the strongest signal, and then applies KNN method.
- Gk method
The main idea of GK method is to calculate the possibility of relative supersaturation of each fingerprint position by Gaussian Kernel Density Estimator (GKDE), and determine the position of TP by averaging the position corresponding to the highest likelihood value.
Triangulation based on Wireless Signal
This method comes from the traditional positioning method, similar to GPS. If we can get the distance or Angle of our undetermined site relative to other base stations, we can find our position by solving multiple equations of intersection columns. Of course, in practice, it is often not to solve equations just enough for positioning. Instead, you take the extra observations and use the least square estimation method to estimate the position. The following figure shows an example. For the three base stations, if we can obtain the time difference between signal transmitting and receiving, then the distance can be obtained through the time difference and the speed of light propagation to obtain positioning.
However, for ordinary WiFi, Bluetooth signals, it is almost impossible to calculate the time difference between the signal being transmitted and the signal being received. Because you have to keep the time on both sides in sync, in order to calculate the time difference between the transmitting time and the receiving time, given the speed of light of C, a little bit of time difference can make a very big difference. Therefore, it is basically impossible to locate through the method of time difference.
So indoor positioning inside the trilateral positioning is how to achieve it?
We know that the location of the target can be solved by taking the distance between the current target and each base station. Since this distance cannot be obtained by the product of time and the speed of light, we can obtain it by another method, that is, the signal attenuation model, as shown in the figure below. We know that in the offline state, wireless signal strength propagates in space and decays with the distance! And that wireless signal strength is measurable to the receiver on your phone! So according to the tested signal strength, can we deduce the distance inversely according to the attenuation model (formula) below, so that we can use three-side positioning again.
