can you share your self -made presets Ring road to me :D —？

Originally posted by @jiaweiwuwu in #3 (comment)

O jaki Ring road Ci chodzi??

• Find or create a vehicle model with LIDAR, camera, odometry and Ackermann steering
• Figure out a way to generate test tracks, e.g. using this package
• Create a Wiki page with instruction for creating tracks and running simulation

Implement node detect_obstacles in package selfie_perception, which will be able to detect positions and dimensions of cuboid obstacles in LIDAR scans.

detect_obstacles

Subscribed topics

scan (sensor_msgs/LaserScan)
Laser scan.

Published topics

obstacles (selfie_msgs/PolygonArray)
List of detected obstacles, represented as rectangles in local coordinate space of the vehicle.

Subscribed transformations

base_link → laser

Implement node change_lane in package selfie_control, which will be responsible for publishing the setpoint for PID controller (#30) depending on which lane the vehicle is supposed to drive at given moment. The setpoint itself is a linear offset (in meters) from the center line. For now a fixed lane width is assumed. By default the vehicle should drive on the right lane.

The node should implement a ChangeLane ActionServer named change_lane. The action handling flow should look like follows:

1. New goal is received
2. If target lane is the same as the one we're currently on, the goal is rejected and we do not proceed further.
3. Otherwise, the goal is accepted, setpoint is changed accordingly and the correct turn indicator is turned on. We keep track of the vehicle's position on the road.
4. Once the vehicle reaches the center of the target lane (within marigin of error), the goal is marked as succeeded and the turn indicator is turned off.

Refer to actionlib package documentation to learn more about actions, especially the section on goal state transitions.

change_lane

Parameters

~lane_width (float)
Assumed lane width (in meters).

~error_margin (float)
Maximum distance from center of the target lane (in meters) until the lane change maneuver is considered complete.

Subscribed topics

position_offset (std_msgs/Float64)
Current distance from the center line (in meters). Positive offset is to the left.

Published topics

target_offset (std_msgs/Float64)
Desirable distance from the center line (in meters), used as setpoint for PID controller. Positive offset is to the left. The value should be set depending on which lane the vehicle is supposed to drive on at given moment.

left_turn_indicator (std_msgs/Bool)
right_turn_indicator (std_msgs/Bool)
Flags used to turn on left and right turn indicator, respectively. Should be set to true for the whole period during which the indicator should be turned on - blinking will be handled in a separate program.

Create node pass_control in package selfie_control which in conjunction with a multiplexer instance will be responsible for passing control to free drive subsystem (#1, #29, #4) after successful execution of the starting procedure (#11).

Along with the node should come a launch file with system configuration, including topic remapping etc.

pass_control

Subscribed topics

attempt_started (std_msgs/Bool)
Published as true at the end of the starting procedure.

Services called

mux/select (topic_tools/MuxSelect)
Used to select subsystem which will be taking control of the vehicle.

Implement node estimate_position in package selfie_path_planner, which given a set of road markings detections will be able to estimate the vehicle's position with respect to the right lane of the road. This node should assume perfect detection independent of special conditions, such as dashed lines, missing markings etc.

estimate_position

Subscribed topics

road_markings (selfie_msgs/RoadMarkings)
Position of road markings, given as lists of points in local coordinate system of the vehicle.

Published topics

position_offset (std_msgs/Float64)
Distance from the center of the right lane (in meters). Positive offset is to the left.

heading_offset (std_msgs/Float64)
Angle between the vehicle's direction and the tangent to the road (in radians), normalized between -π and π. Positive offset is to the left.

Implement node starting_procedure in package selfie_starting_procedure which will be responsible for controlling the vehicle during starting procedure for both dynamic events, according to Basic-Cup Regulations. The node will be started while still in the starting box (section 6.3.3), so it should detect when the gate is opened and then drive to the track. Once it gets there, it should publish a message signaling start of the attempt and then terminate.

If there is more than one node needed for the task, a launch file can be created instead.

starting_procedure

Parameters

~config_file (string)
Path to the configuration file containing homography matrices.

Subscribed topics

image (sensor_msgs/Image)
Camera image.

scan (sensor_msgs/LaserScan)
Laser scan.

odom (nav_msgs/Odometry)
Odometry data.

Published topics

drive (ackermann_msgs/AckermannDriveStamped)
Steering commands for the vehicle.

attempt_started (std_msgs/Bool)
This topic should be published exactly once as true once the vehicle reaches the track.

Subscribed transformations

base_link → laser

Adapt PID controller from Selfie 3.0 to generate steering commands given positional and angular offset, as defined in #29.

Improve on detect_markings node in selfie_perception package so that all conceivable scenarios in the Free Drive (w/o Obstacles) and Parking dynamic event from Basic-Cup are handled properly.

These include:

• missing markings
• passing the starting lane
• other part of the track in the field of view
• artifacts outside the road area, including markings and obstacles
• driving through the parking area
• rural intersections (both with and without right of way)

For more details about road specification refer to Basic-Cup Regulations.

This issue replaces #8, #9, #10.

Modify node estimate_position in package selfie_path_planning so that the position_offset will be relative to the center line - instead of the center of the right lane, as it is at the moment.

Create custom node that will take Image from IDS Camera and pass it to ROS.
Reason: ROS packages did not work with our IDS Camera type

Implement node detect_markings in package selfie_perception, which given a stream of rectified images along with the homography matrix from the image plane to the road plane will be able to detect positions of road markings with respect to the vehicle. For now, we assume all markings are white and none are missing. For more details about road layout refer to Carolo Basic-Cup Regulations, especially points 5.2.1.2 and 5.2.1.3.

The node should follow the specification given below. In addition, the use of markers to visualize position of both lane lines is highly encouraged.

detect_markings

Parameters

~config_file (string)
Path to the configuration file containing homography matrices.

~point_density (float)
Desired distance between subsequent line points, in meters.

~range (float)
How far ahead to look for road markings, in meters.

Subscribed topics

image_rect (sensor_msgs/Image)
Rectified image.

Published topics

road_markings (selfie_msgs/RoadMarkings)
Markings detected from video, given as lists of points in local coordinate system of the vehicle.