Skip to content

Perception#

graph TD
    cmp_sen("Sensing"):::cls_sen
    cmp_loc("Localization"):::cls_loc
    cmp_per("Perception"):::cls_per
    cmp_plan("Planning"):::cls_plan

    msg_img("<font size=2><b>Camera Image</b></font size>
    <font size=1>sensor_msgs/Image</font size>"):::cls_sen

    msg_ldr("<font size=2><b>Lidar Point Cloud</b></font size>
    <font size=1>sensor_msgs/PointCloud2</font size>"):::cls_sen

    msg_lanenet("<font size=2><b>Lanelet2 Map</b></font size>
    <font size=1>autoware_auto_mapping_msgs/HADMapBin</font size>"):::cls_loc

    msg_vks("<font size=2><b>Vehicle Kinematic State</b></font size>
    <font size=1>nav_msgs/Odometry</font size>"):::cls_loc

    msg_obj("<font size=2><b>3D Object Predictions </b></font size>
    <font size=1>autoware_auto_perception_msgs/PredictedObjects</font size>"):::cls_per

    msg_tl("<font size=2><b>Traffic Light Response </b></font size>
    <font size=1>autoware_perception_msgs/TrafficSignalArray</font size>"):::cls_per

    msg_tq("<font size=2><b>Traffic Light Query </b></font size>
    <font size=1>TBD</font size>"):::cls_plan


    cmp_sen --> msg_img --> cmp_per
    cmp_sen --> msg_ldr --> cmp_per
    cmp_per --> msg_obj --> cmp_plan
    cmp_per --> msg_tl --> cmp_plan
    cmp_plan --> msg_tq -->cmp_per

    cmp_loc --> msg_vks --> cmp_per
    cmp_loc --> msg_lanenet --> cmp_per

classDef cmp_sen fill:#F8CECC,stroke:#999,stroke-width:1px;
classDef cls_loc fill:#D5E8D4,stroke:#999,stroke-width:1px;
classDef cls_per fill:#FFF2CC,stroke:#999,stroke-width:1px;
classDef cls_plan fill:#5AB8FF,stroke:#999,stroke-width:1px;

Inputs#

PointCloud#

PointCloud data published by Lidar.

Image#

Image frame captured by camera.

Vehicle kinematic state#

current position of ego, used in traffic signals recognition. See output of Localization.

Lanelet2 Map#

map of the environment. See outputs of Map.

Output#

3D Object Predictions#

3D Objects detected, tracked and predicted by sensor fusing.

Traffic Signals#

traffic signals recognized by object detection model.