LOCATION

Positioning is the most basic and crucial step in the driving navigation. An accurate positioning can effectively improve the accuracy of the road-binding, and can also sense the change of the driving pattern more accurately. Since the project is mainly based on mobile phones for navigation and positioning, Currently used is the built-in sensor data (gyroscope, accelerometer, geomagnetic meter, direction sensor, gravity sensor) and GPS data fusion positioning.

Sensor data required.

• gyroscope(x, y, z).
• accelerometer(x, y, z).
• geomagnetic meter(x, y, z).
• gravity sensor(x, y, z).
• direction sensor(roll, pitch, yaw).
• compass(degree).
• road info(distance to next cross, bearing, road type).
• GPS(lng, lat, alt, accuracy, speed, bearing, t).

Some implement details

• sensor data filter.

• GPS fusion INS under uncoupling system.

Quick start

First make sure gcc and cmake was installed, and include this library into your project.

git clone https://github.com/yyccR/Location.git

Second open your CMakeLists.txt and add these:

include_directories(${PROJECT_SOURCE_DIR}/Location/include/eigen3)

include_directories(${PROJECT_SOURCE_DIR}/Location/math)
add_subdirectory(Location/math)

include_directories(${PROJECT_SOURCE_DIR}/Location/models)
add_subdirectory(Location/models)

include_directories(${PROJECT_SOURCE_DIR}/Location/location)
add_subdirectory(Location/location)

include_directories(${PROJECT_SOURCE_DIR}/Location/sensor)
add_subdirectory(Location/sensor)

include_directories(${PROJECT_SOURCE_DIR}/Location/system)
add_subdirectory(Location/system)

target_link_libraries(${PROJECT_NAME} Location_math)
target_link_libraries(${PROJECT_NAME} Location_models)
target_link_libraries(${PROJECT_NAME} Location_location)
target_link_libraries(${PROJECT_NAME} Location_sensor)
target_link_libraries(${PROJECT_NAME} Location_system)
target_link_libraries(${PROJECT_NAME} Location_test)

final open your main file, and add the test code.

#include <iomanip>
#include <Eigen/Dense>
#include "sensor/GPS.h"
#include "location/Location.h"

using namespace Eigen;
using namespace std;

int main() {

    Location location;
    Vector3d gyro_data_v(0.004263,0.019169,-0.001014);
    Vector3d mag_data_v(-2.313675,-82.446960,-366.183838);
    Vector3d acc_data_v(0.105081,0.108075,9.774973);
    VectorXd gps_data_v(7);
    gps_data_v << 114.174118,22.283789,0.0,0.0,24.0,0.0,1554348968704.665039;
    Vector3d g_data_v(0.094139, 0.107857,9.808955);
    Vector3d ornt_data_v(-0.549866,0.629957,-0.069398);
    Vector3d road_data(1000.0, 0.0, 0);
    location.PredictCurrentPosition(gyro_data_v,acc_data_v,mag_data_v,gps_data_v,g_data_v,ornt_data_v, road_data);
    cout << location.GetGNSSINS().lng << " " << location.GetGNSSINS().lat << endl;
    return 0;
}

if you see the output 114.174 22.2838 that means this library was embedded to your project successfully.

Input data format.

• gyroscope(x, y, z), origin gyroscope data, unit rad/s

• accelerometer(x, y, z), origin accelerometer data, unit m/s²

• geomagnetic meter(x, y, z), origin geomagnetic data, unit μt

• gravity sensor(x, y, z), origin gravity data, unit m/s²

• direction sensor(roll, pitch, yaw), origin sensor data, unit degree

Note that direction sensor doesn't exit actually , the 'sensor data' is computation result from system underlying algorithm.

• compass(degree), origin sensor data, unit degree

• road info(distance to next cross, bearing, road type)

This data is from map data, and if you couldn't search map server data, just fill in all zero (0.0, 0.0, 0.0)

• GPS(lng, lat, alt, accuracy, speed, bearing, t)
  • lng, longitude, double
  • lat, latitude, double
  • alt, altitude, double
  • accuracy, double
  • speed, double
  • bearing, double, unit degree
  • t, timestampe, unit millisecond

More detail tutorial.

• Api calls details
• Sensor data checking
• Impelement details
• Sensor calibration
• Training Stop detection model

TODO

• improve CMake.
• Clean the garbage code.
• Template processing.
• Using smart pointer instead.
• Complete all kinds of documents.
• Add quick start.
• Add more test case.
• Design a suitable pattern.

reference:

1. 《惯性导航》秦永元
2. 《捷联惯性导航技术(第2版 译本)》译者:张天光/王秀萍/王丽霞 作者:DavidH.Titte
3. An efficient orientation filter for inertial and inertial/magnetic sensor arrays
4. Estimation of IMU and MARG orientation using a gradient descent algorithm
5. Direction Cosine Matrix IMU Theory
6. METHODS FOR NON-LINEAR LEAST SQUARES PROBLEMS
7. A Calibration Algorithm for Microelectromechanical Systems Accelerometers in Inertial Navigation Sensors
8. A Calibration Method of Three-axis Magnetic Sensor Based on Ellipsoid Fitting
9. Accuracy Improvement of Low Cost INS/GPS for Land Applications
10. Trajectory preprocessing: Computing with Spatial Trajectories