The codes of embedded system for robotics experiment in summer school. It can implement the functions are: simple serial control, simple ultrasonic obstacle avoidance and following based on face recognition.

1、woodbox_mobile_car_serial_control:
Control the car move by using serial port.

2、woodbox_mobile_car_obstacle_avoidance:
Make the car avoid obstacles by using ultrasonic sensor.

3、woodbox_mobile_car_follow_people:
Make the car follow people by using face recognition from jetson tk1.

1、face_capture:
Capture some face's pictures from camera, convert them to gray images and resize them to default size.

2、face_detection:
Detect faces by using opencv haar cascade classifier.

3、face_train:
Before starting to recognize face, use some face's pictures to train for generating face's model file.

4、face_recognition:
Load face's model file and receive specific faces to start face recognition.

Use following command to get woodbox-mobile-car sources:

$> cd ~/Desktop
$> git clone https://www.github.com/myyerrol/woodbox-mobile-car.git

1、Move the arduino/WoodboxMobileCar floder to your Arduino/libraries. In linux, the default location of arduino libraries is ~/Arduino/libraries, if you change it, please use new path to replace under!

$> cp -r woodbox-mobile-car/arduino/WoodboxMobileCar ~/Arduino/libraries

2、Open the arduino software, and you can find three demo files in the File/Examples/WoodMobileCar option. You can load any of these to test their function.

1、face_capture:
This part of codes use cmake build system. If you want to use IDE to edit or improve these codes, you can start qtcreator and open CMakeLists.txt file to load this project. Of course, you also can build them in the terminal directly by using following command.

$> cd woodbox-mobile-car/jetson_tk1/face_capture
$> mkdir build
$> cd build
$> cmake ..
$> make
$> ./face_capture

2、face_detection:
This part of codes use cmake build system. If you want to use IDE to edit or improve these codes, you can start qtcreator and open CMakeLists.txt file to load this project. Of course, you also can build them in the terminal directly by using following command.

$> cd woodbox-mobile-car/jetson_tk1/face_detection
$> mkdir build
$> cd build
$> cmake ..
$> make
$> ./face_detection

3、face_recognition:
Before building, please replace the paths of face_train.txt to yours!

$> cd woodbox-mobile-car/jetson_tk1/face_recognition
$> make all
$> ./face_train face_train.txt
$> ./face_recognition

Though the time is limited, I still have learned some knowledges about how to use opencv to implement simple face detection and recognition in nvidia jetson tk1. I believe I can do better in the future!