Development Env.

• ROS melodic
• Used Open Source Frameworks & Libraries
  • navigation
  • catographer_ros
    • Modify backpack_2d.lua file
      • num_point_clouds = 1,
  • darknet ros with YOLO
    • Modify darknet_ros.launch file
      • <arg name="image" default="/main_camera/image_raw" />
    • Add yolo_v2_tiny.launch

        <?xml version="1.0" encoding="utf-8"?>
      
        <launch>
      
          <!-- Use YOLOv3 -->
          <arg name="network_param_file"         default="$(find darknet_ros)/config/yolov2-tiny.yaml"/>
          <arg name="image" default="main_camera/image_raw" />
      
      
          <!-- Include main launch file -->
          <include file="$(find darknet_ros)/launch/darknet_ros.launch">
            <arg name="network_param_file"    value="$(arg network_param_file)"/>
            <arg name="image" value="$(arg image)" />
          </include>
      
        </launch>
      
      

  • jetson_camera
  • rplidar_ros
  • pointcloud_to_laserscan
  • usb_cam
• Used Hardware
  • Jetbot AI Advanced (made by yahboom)
  • Jetson nano

---

FSM(Finite State Machine) state

1. IDLE
2. COMMUNICATION
   • Goal point 받기
   • 환자 정보 받기
   • 물품 정보 받기
   • 요청 결과 알리기
   • 미션 수행 요청 받기
3. PATH_PLANNING
4. OPERATION
   • MOVING (Controller)
   • PAUSE
   • DETECTION

---

Mission Flow

secnario

1. IDLE
   • waiting for request from medical team
   • request: goal position
     • goal type
     • x, y axis
2. GO_TO_PATIENT
   • DETECT_OBSTACLES
   • CONTROL
   • COMMUNICATION
3. DETECT_PATIENT
   • detect a patient
   • visual servoing
4. WAIT_FOR_RESPONSE_FROM_PATIENT
5. GO_BACK_HOME

---

Modules

Nodes

1. Mission Planner
2. Cartographer
   • rplidar
   • imu(myahrs+)
3. Darknet_ros
   • jetson_camera
4. Motor_controller
5. Servo_controller

JS (Server)

1. Communicate
   • Patient -> Server
   • Medical -> Server
   • Server -> Robot

---

Pub & Sub

1. Mission Planner

   • Subscriber
     • Communication data
     • Navigation data
     • BoundingBox data
     • Lidar data
   • Publisher
     • Motor cmd data
     • Servo cmd data

2. Cartographer

   • Subscriber
     • Imu data
     • Lidar data
   • Publisher
     • Navigation data

3. Darknet_ros

   • Subscriber
     • Image data
   • Publisher
     • BoundingBox data

4. Motor_controller

   • Subscriber
     • Motor cmd data
     • Navigation data

5. Servo_controller

   • Subscriber
     • Servo cmd data

---

Data

Communication data

• std_msgs/Float32MultiArray ("/rrt/goal_point")

  • data[0] = [0] Go to patient, [1] Go to home
  • data[1] = targetPositionX
  • data[2] = targetPositionY

• std_msgs/Bool ("/rrt/is_taken")

  • data = [False] not yet, [True] yes!!!!!!

Navigation data

• nav_msgs/Odometry

BoundingBox data

• darknet_ros/BoundingBoxs "/darknet_ros/bounding_boxes"
  • Header header
  • Header image_header
  • BoundingBox[] bounding_boxes
    • float64 probability
    • int64 xmin
    • int64 ymin
    • int64 xmax
    • int64 ymax
    • int16 id
    • string Class

Motor cmd data

• ackermann_msgs/AckermannDrive
  • float32 steering_angle
  • float32 steering_angle_velocity
  • float32 speed
  • float32 acceleration
  • float32 jerk

Servo cmd data

• std_msgs/Float32MultiArray
  • data[0] = panAngle
  • data[1] = tiltAngle

Imu data

• sensor_msgs/Imu "/imu/data_raw"

Lidar data

• sensor_msgs/PointCloud2 "/scan"

Image data

• sensor_msgs/Image "/usb_cam/image_raw"