legendtkl Requirements for Outputs
General Guidelines
PDDL Files
- Domain files must define actions that manipulate spatial states.
- Problem files should specify start and goal states clearly.
- Trajectories must adhere to sequential action constraints.
Planner Behavior
- Trajectory generation must complete within specified time.
- If no trajectory exists, return an empty trajectory or failure flag.
PDDL Trajectory Planner
1. Load Domain and Problem
load-trajectory-problem(domain_path, problem_path)
Description:
Loads a PDDL domain and problem file specific to trajectory planning, creating a planning problem instance.
Parameters:
domain_path(str): Path to the PDDL domain file.problem_path(str): Path to the PDDL problem file.
Returns:
trajectory_problem: A trajectory planning problem object.
Example:
trajectory_problem = load_trajectory_problem("domain.pddl", "navigation_task.pddl")
Notes:
- Uses unified_planning.io.PDDLReader.
- Throws an error if parsing is unsuccessful.
2. Generate Trajectory
generate-trajectory(trajectory_problem)
Description: Generates a series of actions (trajectory) for the specified planning problem.
Parameters:
trajectory_problem: A trajectory planning problem instance.
Returns:
trajectory: A list of sequential actions representing the trajectory.
Example:
trajectory = generate_trajectory(trajectory_problem)
Notes:
- Utilizes
unified_planning.shortcuts.TrajectoryPlanner. - If no trajectory can be formed, it returns None.
3. Save Trajectory
save-trajectory(trajectory, output_path)
Description: Writes a trajectory to a file in a defined format suitable for future use.
Parameters:
trajectory: A list of actions representing the trajectory.output_path(str): Path where the trajectory will be saved.
Example:
save_trajectory(trajectory, "navigation_solution.trajectory")
Notes:
- Utilizes a custom file writer for trajectory formats.
- Output is a structured text file.
4. Validate Trajectory
validate-trajectory(trajectory_problem, trajectory)
Description: Validates that the generated trajectory meets the constraints and goals defined in the given problem.
Parameters:
trajectory_problem: The trajectory planning problem.trajectory: The generated trajectory.
Returns:
- bool: True if trajectory is valid, False otherwise.
Example:
is_valid = validate_trajectory(trajectory_problem, trajectory)
Notes:
- Uses
unified_planning.shortcuts.TrajectoryValidator. - Checks action feasibility and goal reachability.
Example Workflow
# Load the problem
trajectory_problem = load_trajectory_problem("domain.pddl", "navigation_task.pddl")
# Generate trajectory
trajectory = generate_trajectory(trajectory_problem)
# Validate trajectory
if not validate_trajectory(trajectory_problem, trajectory):
raise ValueError("Generated trajectory is invalid")
# Save trajectory
save_trajectory(trajectory, "navigation_solution.trajectory")
Notes
- This skill set is designed for effective trajectory planning in dynamic environments.
- Suitable for research in robot navigation and path optimization.