🎮 Nav2 Controllers Extensions

Section

🧭 Navigation - high-level documentation

The pal_nav2_controllers package extends the nav2_controllers package, providing implementations of controller modules which act as plugins of the nav2 base class nav2_core::Controller, also known as Local Planners.

These controller plugins are loaded by the Nav2 Controller Server and are used when calling the /follow_path.

PID Controller

The PID Controller implements a motion planning algorithm that can be used for path following or for pose tracking.

It implements a simple PID Control

strategy to minimize the distance of the robot from any given path or any give goal.

Apart from being used as a Nav2 Controller Plugin, it can also be used within the Target Navigation framework as a Target Tracker.

It accepts a nav_msgs::msg::Path with just one Pose and reaches it by minimizing at the same time the x, y, and theta error from the given Pose.

Alternatively, this functionality can also be triggered by sending an action goal to the /follow_path.

Send a goal from the command-line

  ros2 action send_goal /follow_path nav2_msgs/action/FollowPath "path:
     header:
        stamp:
           sec: 0
           nanosec: 0
        frame_id: 'base_footprint'
     poses:
       - header:
           stamp:
             sec: 0
             nanosec: 0
           frame_id: 'base_footprint'
         pose:
           position:
             x: 0.0
             y: 0.0
             z: 0.0
           orientation:
             x: 0.0
             y: 0.0
             z: 0.0
             w: 1.0
     controller_id: 'TrackPose'
     goal_checker_id: ''"

You can set the usage of the PID Controller by setting the controller_id field to 'TrackPose' and change the pose.position and the pose.orientation values to the desired pose. Furthermore, by setting the frame_id to the base_footprint, this works as a Cartesian Controller that moves the robot to a goal expressed relative to its base frame. In this configuration, it can also be used for Instruction-based Navigation (e.g. move three meters forward, then turn right).

Obstacle Avoidance

This controller implements a basic collision detection algorithm that before sending a velocity commands to the robot, simulates its consequences (e.g. where will the robot be after this command). If a collision is detected, the robot is stopped.

Configuration

Parameter	Description
simulate_ahead_time	The time (in seconds) to project the velocity when looking for collisions.
max_linear_vel	Maximum linear velocity (m/s) allowed by the controller.
max_angular_vel	Maximum angular velocity (rad/s) allowed by the controller.
kp_gain	Proportional Gain of the PID controller.
ki_gain	Integral Gain of the PID controller.
kd_gain	Derivative Gain of the PID controller.
antiwindup	Whether to enable the antiwindup.
integration_clamp	Max Integral error allowed.
step_size	Sampling size of the path. Lower values correspond to stricter path following.
motion_model	Desired motion model used for motion planning. Options are: 'DiffDrive', 'OmniDrive'.
transform_tolerance	Time tolerance for data transformations with TF (in seconds).

All the parameters above are dynamic, and can be changed at runtime producing a change in the behavior of the controller.