CN104808231B - Unmanned plane localization method based on GPS Yu light stream Data Fusion of Sensor - Google Patents

Abstract

The present invention relates to a kind of unmanned plane localization method, to realize unmanned plane autonomous positioning, therefore, the present invention is adopted the technical scheme that, the unmanned plane localization method based on GPS Yu light stream Data Fusion of Sensor comprises the following steps：The velocity information and velocity information of unmanned plane are obtained using the light stream sensor installed in four rotor wing unmanned aerial vehicle bottoms, the positional information of unmanned plane is obtained using GPS and judges the reliability of positional information, above- mentioned information is merged using complementary filter, in the alignment system of unmanned plane.Present invention is mainly applied to unmanned plane positioning.

Description

UAV positioning method based on GPS and optical flow sensor data fusion

technical field

The invention relates to a positioning method for an unmanned aerial vehicle, in particular to an autonomous positioning method for an unmanned aerial vehicle based on fusion of GPS and optical flow sensor data.

Background technique

The positioning problem of the UAV mainly refers to the use of its own sensors to determine the position and attitude information of the UAV relative to the inertial coordinate system in the flight environment. Accurate pose estimation is the premise and basis for complex flight tasks such as safe flight, trajectory planning, and target tracking of quadrotor UAVs.

Currently widely used UAV navigation system is mainly based on GPS positioning method, but its positioning accuracy is low. In recent years, although the UAV positioning system based on the optical flow method has high positioning accuracy, its positioning results have a large position drift with time.

Researchers from the National University of Singapore used the algorithm combining Kalman filter and complementary filter and the H∞ robust filter algorithm to successfully fuse the position information and velocity information of the UAV obtained by GPS, overcoming the inconsistency of the position information obtained by GPS. Continuity, to obtain relatively accurate and smooth UAV position information, but due to the limitation of GPS data output accuracy, its positioning accuracy is about ±1 meter (Conference: IEEE International Conference on Control and Automation; Author: Yun B, Peng K , Chen B M; publication year: 2007; article title: Enhancement ofGPS Signals for Automatic Control of a UAV Helicopter System; page number: 1185-1189) (Journal: Transactions of the Institute of Measurement and Control; Author: Yun B, CaiG, Chen B M, Year and Month of Publication: 2011; Article Title: GPS Signal Enhancement and Attitude Determination for a Mini and Low-cost Unmanned Aerial Vehicle, Pages: 665–682).

Researchers from the Swiss Federal Institute of Technology in Zurich used the PX4FLOW optical flow sensor as a position measurement unit and used it in the positioning control system of indoor and outdoor drones. Although the research institution realized the indoor trajectory tracking experiment based on the optical flow method, the It can be seen that after the long-distance integration of the speed information of the UAV obtained by the optical flow sensor, the maximum positioning error is greater than 0.5 meters in the process of tracking two consecutive rectangular trajectories with a side length of about 3 meters. There is a large drift in its position information during the process (Conference: IEEE International Conference on Robotics and Automation; Author: Honegger D, Meier L, Tanskanen P; Year of publication: 2013; Article title: AnOpen Source and Open Hardware Embedded Metric Optical Flow CMOS Camera for Indoor and Outdoor Applications; pp. 1736-1741).

In addition, researchers from the Swiss Federal Institute of Technology in Zurich also used the optical flow method as an auxiliary position measurement unit for the positioning control system of the drone, but due to the large position drift caused by the integral effect of the optical flow method during the long-distance movement , there is a large positioning error between its two consecutive weeks of rectangular trajectories (Conference: IEEE/RSJ InternationalConference on Intelligent Robots and Systems; Author: Fraundorfer F, Heng L, Honegger D; Year of publication: 2012; Article title: Vision- based Autonomous Mapping and Exploration Using a Quadrotor MAV; pp. 4557-4564).

Contents of the invention

In order to overcome the deficiencies in the prior art and realize the autonomous positioning of the UAV, the technical solution adopted by the present invention is a method for positioning the UAV based on the fusion of GPS and optical flow sensor data, including the following steps:

Use the optical flow sensor installed at the bottom of the quadrotor UAV to obtain the speed information and speed information of the UAV, use GPS to obtain the position information of the UAV and judge the reliability of the position information, and use complementary filters to fuse the above information , used in the positioning system of the UAV.

The speed information and speed information processing of the speed information obtained by the optical flow sensor of the unmanned aerial vehicle are:

The optical flow method directly obtains the horizontal velocity information of the UAV, so it needs to obtain its position information through integral calculation. It filters out.

The position information obtained by the GPS method and the speed information obtained by the optical flow method are simplified into the following form:

y _p ＝p+μ _p

y _v ＝v+μ _v

Here, p and v are the real position and velocity measurements obtained by GPS method and optical flow method respectively, and μ _p and μ _v are the measurement noise of position information and velocity information respectively, both of which are constant values.

Described GPS obtains the position information of unmanned aerial vehicle and judges the reliability of position information is:

Due to the instability of GPS information, it is necessary to detect GPS information. If the number of satellites used in the positioning process is less than 5, it is considered that the position information of the UAV obtained by this method is not accurate enough.

Using a complementary filter to fuse the above information is specifically: the output of the measurement results after passing through the complementary filter The frequency domain expression of is written as follows:

Where s is the Laplacian operator, T(s) is the first-order low-pass filter, S(s) is the first-order high-pass filter, T(s)+S(s)=1, y _p (s) means The sensor obtains position information in the frequency domain; the complementary filter controller adopts proportional feedback, that is, C(s)=k; in this case, the dynamic equation of the closed-loop system is

At this time, the filter frequency domain expression form is with The cutoff frequency of the designed filter is In the high frequency band greater than f _T , the optical flow sensor data plays a major role in the fusion result, and in the low frequency band smaller than f _T , the GPS data plays a major role in the fusion result, and k is a constant.

Compared with prior art, technical characteristic and effect of the present invention:

The invention uses a complementary filtering algorithm to fuse the position information of the UAV obtained by GPS and optical flow method, so as to realize the acquisition of high-precision and accurate position information of the UAV in a relatively long time range, and meet the requirements of long-term autonomous flight control of the UAV. need.

Description of drawings

Fig. 1 is the structural block diagram of complementary filter that the present invention adopts;

Figure 2 is a schematic diagram of the trajectory of the UAV in the horizontal direction, in which,

Figure 2a is the trajectory of the UAV in the horizontal direction acquired by GPS;

Figure 2b is the horizontal motion trajectory of the UAV acquired by the optical flow sensor;

Figure 2c is the horizontal movement trajectory of the UAV obtained by using the data fusion algorithm.

detailed description

The technical problem to be solved by the present invention is to provide an autonomous positioning method for unmanned aerial vehicles based on fusion of GPS and optical flow sensor data, so as to realize accurate and drift-free positioning of unmanned aerial vehicles in outdoor environments.

The technical scheme adopted in the present invention is: adopt the method for GPS and optical flow sensor data fusion to be used in the positioning system of unmanned aerial vehicle, comprise the following steps:

Use the optical flow sensor installed at the bottom of the quadrotor UAV to obtain the speed information and speed information of the UAV, use GPS to obtain the position information of the UAV and judge the reliability of the position information, and use complementary filters to fuse the above data , used in the positioning system of the UAV.

The speed information and speed information processing of the speed information obtained by the optical flow sensor of the unmanned aerial vehicle are:

The optical flow method directly obtains the horizontal velocity information of the UAV, so it needs to obtain its position information through integral calculation. It filters out.

Described GPS obtains the position information of unmanned aerial vehicle and judges the reliability of position information is:

Due to the instability of GPS information, its signal strength will be affected by solar activities, etc., so it is necessary to detect GPS information in this paper. If the number of satellites used in the positioning process is less than 5, it is considered that the method obtained The location information of the drone is not accurate enough;

Described adopting complementary filter to carry out data fusion is:

The complementary filtering algorithm is used to fuse the position information of the UAV obtained by GPS and optical flow method, so as to obtain the high-precision and accurate position information of the UAV in a long time range, and meet the needs of long-term autonomous flight control of the UAV. The following will mainly introduce the data fusion algorithm based on complementary filter.

Since the positioning algorithm based on GPS and optical flow method can only directly or indirectly obtain the horizontal position information of the UAV, only the position information fusion method of the UAV in the horizontal direction is considered in the present invention.

In an ideal state, the UAV’s position information p=[xy] ^T and its corresponding speed information v=[v _x v _y ] ^T have the following relationship:

Among them, x and y are the position information of the UAV in the X and Y directions respectively, v _x and _vy are the velocity information of the UAV in the X and Y directions respectively, and the left side of the equation is the first derivative of p.

However, in the actual measurement process, due to the sensor's own accuracy and the existence of external interference, the measurement results often contain a large number of noise signals and interference information. Here, the position information obtained by the GPS method and the speed information obtained by the optical flow method are simplified as In the following form:

y _p ＝p+μ _p

y _v ＝v+μ _v

Here, p and v are the real position and velocity measurements obtained by GPS method and optical flow method respectively, and μ _p and μ _v are the measurement noise of position information and velocity information respectively, both of which are constant values.

The structural block diagram of the complementary filtering algorithm is shown in Figure 1. According to the structural block diagram of the system, the measurement results can be output after passing through the complementary filter The frequency domain expression of is written as follows:

Where s is the Laplacian operator, T(s) is the first-order low-pass filter, S(s) is the first-order high-pass filter, T(s)+S(s)=1, y _p (s) means The sensor obtains position information in the frequency domain.

In the actual control system, according to the low-pass characteristics of y _p (s) and the high-pass characteristics of y _v (s), the complementary filter controller is designed as proportional feedback, that is, C(s)=k; in this case, it can be The dynamic equation of the closed-loop system is obtained as

At this time, the filter frequency domain expression form is with The cutoff frequency of the designed filter is In the high frequency band greater than f _T , the optical flow sensor data plays a major role in the fusion result, and in the low frequency band smaller than f _T , the GPS data plays a major role in the fusion result, and k is a constant.

The autonomous positioning method of the drone based on the fusion of GPS and optical flow sensor data of the present invention will be described in detail below in combination with the embodiments and the accompanying drawings.

The present invention comprehensively considers the advantages of the positioning method based on GPS and the positioning method based on the optical flow method, fully utilizes the advantages of long-term long-distance positioning of GPS and high positioning accuracy of the optical flow method in a short time, and adopts a complementary filtering method to combine the two positioning methods The algorithm is fused, and the corresponding experimental verification is carried out in an outdoor environment with a handheld quadrotor UAV, and the precise positioning function of the quadrotor UAV based on multi-sensor data fusion is realized.

The present invention is based on the unmanned aerial vehicle autonomous positioning method of GPS and optical flow sensor data fusion, comprises the following steps:

1) Use the optical flow sensor to obtain the speed information of the UAV and process the speed information:

The optical flow method directly obtains the horizontal velocity information of the UAV, so it needs to obtain its position information through integral calculation. It filters out.

2) Use GPS to obtain the location information of the drone and judge the reliability of the location information:

Due to the instability of GPS information, its signal strength will be affected by solar activities, etc., so it is necessary to detect GPS information in this paper. If the number of satellites used in the positioning process is less than 5, it is considered that the method obtained The location information of the drone is not accurate enough.

3) Data fusion using complementary filters is:

The complementary filtering algorithm is used to fuse the position information of the UAV obtained by GPS and optical flow method, so as to obtain the high-precision and accurate position information of the UAV in a long time range, and meet the needs of long-term autonomous flight control of the UAV. The following will mainly introduce the data fusion algorithm based on complementary filter.

Since the positioning algorithm based on GPS and optical flow method can only directly or indirectly obtain the horizontal position information of the UAV, this chapter only considers the position information fusion method of the UAV in the horizontal direction.

In an ideal state, the UAV’s position information p=[xy] ^T and its corresponding speed information v=[v _x v _y ] ^T have the following relationship:

Among them, x and y are the position information of the UAV in the X and Y directions respectively, v _x and v _y are the velocity information of the UAV in the X and Y directions respectively, and the left side of the equation is the first derivative of p.

However, in the actual measurement process, due to the sensor's own accuracy and the existence of external interference, the measurement results often contain a large number of noise signals and interference information. Here, the position information obtained by the GPS method and the speed information obtained by the optical flow method are simplified as In the following form:

y _p ＝p+μ _p

y _v ＝v+μ _v

Here, p and v are the real position and velocity measurements obtained by GPS method and optical flow method respectively, and μ _p and μ _v are the measurement noise of position information and velocity information respectively, both of which are constant values.

The structural block diagram of the complementary filtering algorithm is shown in Figure 1. According to the structural block diagram of the system, the measurement results can be output after passing through the complementary filter The frequency domain expression of is written as follows:

Where s is the Laplacian operator, T(s) is the first-order low-pass filter, S(s) is the first-order high-pass filter, T(s)+S(s)=1, y _p (s) means The sensor obtains position information in the frequency domain.

In the actual control system, according to the low-pass characteristic of y _p (s) and the high-pass characteristic of y _v (s), the complementary filter controller is designed as proportional feedback, that is, C(s)=k; in this case, it can be The dynamic equation of the closed-loop system is obtained as

At this time, the filter frequency domain expression form is with The cutoff frequency of the designed filter is In the high-frequency band greater than _fT , the optical flow sensor data plays a major role in the fusion result, and in the low-frequency band smaller than _fT , the GPS data plays a major role in the fusion result.

Specific examples are given below:

1. System hardware connection and configuration

The vision-based autonomous flight control method of the quadrotor drone of the present invention adopts a flight control structure based on an embedded architecture, and the built experimental platform includes a quadrotor drone body, a ground station, a remote controller, and the like. Among them, the quadrotor drone is equipped with an embedded computer (the computer is embedded with an Intel Core i3 dual-core processor with a main frequency of 1.8GHz), an airborne PX4FLOW optical flow sensor, GPS and a flight controller (including an inertial navigation unit and a barometer module. Wait). The ground station includes a laptop computer with a Linux operating system, which is used for onboard program startup and remote monitoring. The platform can be manually taken off and landed through the remote control, and can be switched to manual mode in case of an accident to ensure the safety of the experiment.

2. Flight experiment results

In this embodiment, multiple groups of flight control experiments are carried out on the above-mentioned experimental platform, and the flight experiment environment is an outdoor campus environment. The control goal is to realize the precise and drift-free positioning function of the quadrotor UAV. In the experiment, the movement trajectory is about a rectangle of 85 meters * 135 meters.

The data curve of the UAV during the outdoor handheld experiment is shown in Figure 2. Among them, the starting point of the UAV in the figure is (0,0), the figure a is the horizontal position track of the drone obtained by using GPS, the figure b is the position track obtained by using the optical flow sensor, and the figure c is the position track obtained by using the proposed paper. The UAV position track obtained by the data fusion algorithm. It can be seen from the figure that at point (0,100), due to the poor GPS signal, there are large fluctuations in the position information of the UAV obtained by using GPS at this time, while the fluctuation of the position information of the fused UAV is small. Furthermore, the effectiveness of the fusion algorithm in this paper is proved. In addition, the starting point and ending point of the UAV position information obtained by GPS basically coincide, and the error is within 5 meters; while the UAV position information obtained by the optical flow method has a large position drift in a long period of time, which is different from the initial The point error is about 29 meters; while using the fusion algorithm proposed in this paper, the positioning error is about 11 meters, which proves the effectiveness of the fusion algorithm in this paper.

Claims (2)

1. a kind of unmanned plane localization method based on GPS Yu light stream Data Fusion of Sensor, it is characterized in that, comprise the following steps： The velocity information of unmanned plane is obtained using the light stream sensor installed in four rotor wing unmanned aerial vehicle bottoms, unmanned plane is obtained using GPS Positional information and judge the reliability of positional information, the unmanned plane that will be obtained using complementary filter using light stream sensor Velocity information, merged using the positional information of the GPS unmanned planes obtained, in the alignment system of unmanned plane；Described GPS obtains the positional information of unmanned plane and judges that the reliability of positional information is：There is unstability due to GPS information, it is necessary to GPS information is detected, if be less than 5 stars for the number of satellite in position fixing process, then it is assumed that the unmanned seat in the plane that this method is obtained Confidence breath is not accurate enough；

The nothing that the use complementary filter is obtained by the velocity information of the unmanned plane obtained using light stream sensor, using GPS Man-machine positional information is merged specially：Measurement result is passed through into the output after complementary filterFrequency-domain expression write Into following form：

$\begin{array}{c}\hat{p}\left(s\right)=\frac{C\left(s\right)}{s+C\left(s\right)}{y}_p\left(s\right)+\frac{s}{C\left(s\right)+s}\frac{y_v\left(s\right)}{s}\\ =T\left(s\right)y_p\left(s\right)+S\left(s\right)\frac{y_v\left(s\right)}{s}\end{array},$

Wherein s is Laplace operator, and T (s) is first-order low-pass ripple, and S (s) is single order high-pass filtering, T (s)+S (s)=1, y_p (s) represent to obtain the positional information in frequency domain by sensor；Complementary filter controller adoption rate feeds back, i.e. C (s)=k；Adopt Following form is reduced to the positional information of GPS methods acquisition and by the velocity information that optical flow method is obtained：

$\begin{array}{c}{y}_p=p+{\mathrm{\μ}}_p\\ y_v=v+{\mathrm{\μ}}_v\end{array},$

Here p, v is respectively the actual position and velocity measurement obtained using GPS methods and optical flow method, μ_pAnd μ_vRespectively position Confidence ceases the measurement noise with velocity information, is steady state value；In this case, the dynamical equation for obtaining closed-loop system is：

$\stackrel{\·}{\hat{p}}=y_v+k(y_p-\hat{p}).$

Now filter frequency domain expression-form isWithThe filter cutoff frequency now designed isMore than f_THigh band, light stream sensing data plays a major role to fusion results, less than f_TLow-frequency range, Gps data plays a major role to fusion results, and k is constant.

2. the unmanned plane localization method as claimed in claim 1 based on GPS Yu light stream Data Fusion of Sensor, it is characterized in that, The velocity information of described utilization light stream sensor acquisition unmanned plane is handled：What optical flow method was directly obtained is the level of unmanned plane Velocity information, so need to obtain its positional information by integral operation, simultaneously as other can not survey the influence of factor, Of short duration mistake measurement is had during tachometric survey, it is necessary to artificially be filtered.