CN104266648A - Indoor location system based on Android platform MARG sensor - Google Patents

Abstract

The invention discloses an indoor location method and system based on an Android platform MARG sensor, belonging to the technical field of location and navigation. The MARG sensor comprises a three-axis magnetometer, a three-axis gyroscope and a three-axis accelerometer. The indoor location system mainly comprises a data acquisition unit, a data processing unit and a display unit. When in location, an Android platform is held in hand by a pedestrian and the location information is acquired through the data acquisition unit, the data processing unit and the display unit. The noise filtration is performed on sensor data by utilizing an EKF (extended Kalman filter) algorithm so as to improve the location precision. In places with weak GPS (global positioning system) signal or without GPS signal, high-precision location can be achieved by virtue of the system. The indoor location system is applicable to the Android platform with the MARG sensor, and is good in real-time property, high in portability, high in location precision and convenient to use, and additional equipment is not needed.

Description

The indoor locating system of Android platform MARG sensor

Technical field

The present invention relates to indoor locating system, particularly relate to a kind of indoor orientation method based on Android platform and system.

Background technology

At present, location technology mainly comprises satnav, network based positioning and some other location technology.Satnav is mainly used in outdoor positioning and navigation, and in place that is indoor or that blocked by buildings, satnav precise decreasing even cannot be located.Network based positioning mainly adopts the methods such as TDOA, AOA, is subject to the impact of multipath and decline, and its positioning precision is low.

Location technology based on indoor comprises WLAN, bluetooth etc.Mainly based on methods such as fingerprint location, RSSI location.Adopt WLAN to carry out indoor positioning, need strictly to control WLAN node emissive power, its node also needs to rearrange, and front current cost is higher, safeguards more complicated.

Indoor positioning technologies based on micro-inertia sensor just progressively develops, and the existing localization method based on inertial sensor is mainly based on particular device, and this equipment is fixed on ankle place, and it wears inconvenience, and cost is higher.

Summary of the invention

In order to overcome above-mentioned the deficiencies in the prior art, the invention provides a kind of indoor orientation method based on Android platform MARG sensor and system.

To achieve these goals, the technical solution adopted in the present invention is: to carry the hardware platform of Android device (great majority are mobile phone) as system of MARG sensor.Hand-held Android device when pedestrian locates, the Z axis in MARG sensor sensing axle is perpendicular to ground level.Data acquisition unit carries out sensor data acquisition by the frequency of 50Hz; Data processing unit carries out noise filtering to sensor raw data, gyro zero correction, magnetic interference elimination, attitude algorithm, gait differentiation, step-size estimation, the elements of a fix resolve; The locating information that display unit utilizes data processing unit to pass back carries out trajectory reproducing, and shows attitude information in real time, positional information.

Compared with prior art, the invention has the beneficial effects as follows: (1) utilizes existing Android platform (such as Android phone, flat board etc.) of carrying MARG sensor as system support, has saved system cost, easy to use; (2) under this method overcomes complex environment, magnetic interference is on the impact of positioning precision; (3) this method adopts the method for BP neural network to carry out pedestrian's step-size estimation, and step-size estimation precision is high; (4) this method realizes the full independent navigation of pedestrian, and real-time track is drawn.

Accompanying drawing explanation

In order to make the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail, wherein:

Fig. 1 is indoor orientation method structural drawing.

Fig. 2 is pedestrian's reckoning algorithm block diagram.

Fig. 3 is the hand-held attitude of Android platform when pedestrian locates and MARG sensor sensing direction of principal axis figure.

Fig. 4 is indoor locating system Organization Chart.

Embodiment

Below in conjunction with accompanying drawing, the present invention is further described.

(1) native system is the indoor locating system integrating hardware and software, and based on Android hardware platform, Android operation system is that positioning software runs core.Fig. 1 gives systems approach structure.

(2) indoor locating system based on Android platform is using the Android platform (referring generally to mobile phone or panel computer) with MARG sensor as hardware support, and 1. data acquisition unit forms primarily of orthogonal three accelerometers of sensitive axes, three gyroscopes and three magnetic sensor combination.Wherein, the X-axis of each sensor, Y-axis, Z-direction are identical.Pedestrian walk location time, Z axis is perpendicular to ground level, and X-axis is vertical with Z axis, and left, Z axis and X, Y-axis are vertical, formation right hand rectangular coordinate system for level.1. data acquisition unit is that 50Hz carries out data sampling to sensor according to frequency.

(3) in fig. 2, data processing unit 2. independent operating at the bottom of Android operation system, by the process to raw data, 3. provide locating information to display unit.

(4) acceleration filter unit 1 pair of acceleration X, Y, Z axis raw data carries out medium filtering.By partially asking for gyro zero time static, and corrected by the inclined correcting unit 2 of gyro zero.

(5) magnetic interference detecting unit 3 judge ought for the previous period in the variance of magnetometer modulus value and magnetometer modulus value and local standard magnetic field intensity carry out Interference Detection, and the covariance matrix revised in EKF algorithm, in real time acceleration, magnetometer, gyrostatic weight in adjustment EKF algorithm.

(6) parameter that obtains according to step (5) of pedestrian's attitude algorithm unit 5, adopt expanded Kalman filtration algorithm (EKF) to carry out fused filtering to acceleration, magnetometer, gyro data, and adopt the attitude information of hypercomplex number attitude algorithm algorithm real-time resolving current time.

(7) pedestrian's brief acceleration modulus value Changing Pattern of walking presents cyclical variation, utilizes this cyclical variation and eigenwert to carry out gait detection.3-axis acceleration modulus value is obtained by formula (2), a in formula _x, a _yand a _zbe respectively the data that accelerometer three axle exports.

$\mathrm{Acc}\_\mathrm{norm}=\sqrt{{a_x}^2+{a_y}^2+{a_z}^2}---\left(1\right)$

Utilize wave digital lowpass filter to carry out filtering to acceleration modulus value, obtain good single peak curve map, eliminate slight jitter according to threshold value and obtain good gait information.

(8) step-size estimation unit 6 adopts classical step-size estimation model and BP neural network, establishes the nonlinear model of pedestrian's height, cadence and stride.This nonlinear model comprises the input layer after two standardization and an output layer neuron, the neuronic quantity of hidden layer and the sample number of training and the neuronal quantity of input layer proportional, and to determine in network training process.

(9) coordinate calculating unit 8 adopts the short pedestrian's reckoning algorithm of essence to carry out coordinate and location compute.Its solution formula is as formula (2).

$\left\{\begin{array}{c}{x}_k=x_0+\underset{i=1}{\overset{k}{\mathrm{\Σ}}}{d}_i\cos {\mathrm{\θ}}_i\\ y_k=y_0+\underset{i=1}{\overset{k}{\mathrm{\Σ}}}{d}_i\sin {\mathrm{\θ}}_i\end{array}\right.---\left(2\right)$

(10) 2. obtain pedestrian by data processing unit to locate and attitude information, display unit 3. upon receiving this information, carries out the display of track drafting and other information, returns step (4) simultaneously.

Claims (3)

1., based on an indoor locating system for Android platform MARG sensor, it is characterized in that: comprise MARG sensor data acquisition unit 1., data processing unit 2., display unit 3..

2. 1. the MARG sensor data acquisition unit described in comprises three axle magnetometers, three-axis gyroscope and three axis accelerometer, and the X, Y, Z axis of each sensor is vertical mutually, and the X, Y, Z axis direction of three sensors is all identical; 2. described data processing unit is carry out data processing to 9 dimension data that 1. data acquisition unit spreads out of, and comprises medium filtering, the differentiation of pedestrian's gait, step-size estimation, EKF, hypercomplex number solve attitude angle; 3. described display unit is carry out trajectory reproducing and display to the locating information that 2. data processing unit exports.

3. the system based on claim 1 carries out the method for indoor positioning: it is characterized in that comprising the steps: that (1) is by 1. hand-held for the data acquisition unit in the indoor locating system of described MARG sensor and fixed pose, Z week in three sensitive axes in making data acquisition unit 1., X-axis level left perpendicular to ground level; (2) 2. data processing unit carries out medium filtering to the sensing data that data acquisition unit obtains, and removes the data of sensor kick; (3), when pedestrian is static, the data that 2. data processing unit exports step (2) are carried out current gyro zero and are partially asked for and correct; (4) 2. data processing unit carries out variance threshold values judgement to the magnetometer data that step (3) exports, if exceed pre-determined threshold, revise data processing unit 2. in correlation parameter, eliminate for magnetic interference; (5) 2. data processing unit utilizes the parameter that step (4) obtains, and carries out fused filtering to acceleration, magnetometer, gyrostatic data; (6) 2. data processing unit utilizes the data that step (5) exports, and solves real-time attitude information by Quaternion Method; (7) acceleration information that 2. data processing unit utilizes step (2) to export carries out modulus value to be asked for, and carries out gait detection by modulus value rule; (8) 2. data processing unit utilizes cadence, stride and the height that step (7) exports, and carries out step-size estimation and solve locating information by BP neural network; (9) locating information exported step (8) 3. shows by display unit.