Bluetooth positioning problems encountered and Solutions
Foreword
He went to the graduation year quarter.
Bluetooth positioning is also done quickly off more than a year, the future may not do this one, and take advantage of new memories but also when some of the problems encountered little sum up, then help a friend take a little detour less.
Supplementary little knowledge
ibeacon and Bluetooth
Ibeacon is similar to a previous low-power Bluetooth technology, Bluetooth technology works. iBeacon is a new feature Apple's September 2013 release of mobile devices equipped with the OS on (iOS7). (From Baidu Encyclopedia)
currently do locate Bluetooth are in fact based on the ibeacon.
Meaning algorithm
In theory, known not in the same line three point coordinates, distance and anchor points are three points that we can be obtained directly anchor point.
What you will also need the algorithm to use it? Since the positioning jitters Bluetooth (Bluetooth with the Bluetooth transmitter and receiver as well as the quality of the environment of the terminal), Bluetooth beacons calculated distance to the anchor point is often very inaccurate, it is impossible to directly determine the location of points, needs to be positioned by the fuzzy algorithm.
Mainstream algorithm
Mainly the three-point positioning method and fingerprinting.
Three positioning method: This either-three have three points. Three-point positioning method was originally known coordinates of three points and three points away to an unknown point, method of seeking the unknown point, but because of the practical application of three-point distance is not precise distance, we can only find a vague solution, three-point method and will have a wide range of variants, of course, can be collectively referred to as three-point location method. Not really refers to three three-point, practical application can be used more, but Bluetooth is commonly used in this area 3 to 5:00, there are many specific implementations, such as weight, least square method, average method. According to a particular implementation can be divided into a number of ways.
Fingerprinting: fingerprinting is the first collection of Bluetooth signal strength in different locations stored in the database, to find the most likely position by comparing the database data and signal acquisition.
Two methods have advantages and disadvantages, first of all fingerprinting accuracy is better, faster computing speed three-point method. Fingerprinting preliminary data to be collected, it is very troublesome.
Problems and Solutions
First statement, but there are ways to solve my personal thinking, does not provide specific practices, may be interested in the comments.
Bluetooth problems
First, the cost is not cheap, a complete Bluetooth device probably between 40 ~ 60RMB, quality is uneven. The strength of their own friend suggested, I believe the cost is very substantial.
Sure most people will choose to buy ready-made Bluetooth devices, after all development takes time. A good Bluetooth this time will help you save a lot of time here very important point is to shop around , not necessarily big name manufacturers like equipment, personal experience.
How Bluetooth qualification, acquisition of signal fluctuations in long enough to observe the signal. Bluetooth locate the most important thing is the stability of the signal, the higher stability, more precise positioning.
Algorithmic problems
Bluetooth devices no matter how good, will appear more or less signal distortion conditions. This time we need to design a filter to filter the error signal. Devaluation of the filter is a Kalman filter, followed by an averaging filter.
If using a three-point location method, then theoretically, the average signal required to select the strongest beacon points three calculations do, however, there will be a case in which the selected three beacon points is not exactly positioned actual distance point last three beacon points . The actual application process, probably due to obstruction or signal distortion, resulting in severe signal fading or some abnormal enhancement (multipath effect), resulting in not entirely correct fix our direct calculation, this time to do anti-noise preparation .
And if it is fingerprinting, then the discrimination between different anchor points requires a higher, more accurate positioning of the higher degree of distinction. For example, A Bluetooth, Bluetooth we stand and A 1 m in front of the rear position of 1 m, measured Bluetooth signal A is basically the same, the algorithm does not know at this time in front of or behind us A, this time If we add Bluetooth signals B, B Bluetooth Bluetooth in front of just two in the a position is inconsistent, then we compare the Bluetooth signal through B, and will be able to know that I am in front of or behind the a's. This is the signal collected some tips. However, this time there is a difficulty for fingerprinting, that is, different signal acquisition device will be inconsistent, the device 1 is 5.5 collect ten minutes, ten minutes collecting device 2 may be 6.5. This time need to consider adaptation of equipment problems, or adaptive algorithm problem .
Device Adaptation
Bluetooth devices require adaptation, power, frequency, power consumption. The stronger the power frequency, the better the receiving device, the more precise positioning, the greater the relative power; conversely poor reception apparatus, positional deviation, but the low power consumption.
to sum up
Bluetooth positioning is actually an optimization problem, make trade-offs in a variety of resource constraints, to find an optimal balance point.
