slam_toolboxのパラメータ調整

[shunki1006]

slam_toolboxのパラメータ調整時に, 利用した各パラメータの説明を示します.
以下のリンクも必要に応じて合わせて参照していただけますと幸いです.
・https://github.com/SteveMacenski/slam_toolbox/blob/foxy-devel/README.md#configuration
・#64

1. Solver Plugin Parameters:

• solver_plugin: Specifies the plugin used for solving the SLAM problem. solver_plugins::CeresSolver indicates the use of the Ceres Solver, a popular library for modeling and solving large, complicated optimization problems.
• ceres_linear_solver: Defines the linear solver method for Ceres. SPARSE_NORMAL_CHOLESKY is a method for solving sparse linear systems.
• ceres_preconditioner: Sets the preconditioner for iterative solvers in Ceres. SCHUR_JACOBI is a type of preconditioner used in optimization.
• ceres_trust_strategy: Determines the trust region strategy. LEVENBERG_MARQUARDT is an algorithm that helps in solving nonlinear least squares problems.
• ceres_dogleg_type: Specifies the type of Dogleg strategy, a numerical optimization method. TRADITIONAL_DOGLEG is a standard approach in this category.
• ceres_loss_function: The loss function used in optimization. None means no specific loss function is applied, using a standard least-squares approach.

2. ROS Parameters:

• odom_frame: The name of the odometry coordinate frame.
• map_frame: The name of the map coordinate frame.
• base_frame: The name of the base coordinate frame, typically attached to the moving robot or device.
• scan_topic: The ROS topic where laser scan data is published. /merged_scan indicates a topic name for merged scans from multiple sources.
• mode: Operating mode of the SLAM toolbox, can be either mapping or localization.

3. Additional Parameters:

• debug_logging: Enables or disables debug logging. false means no debug logging.
• throttle_scans: Determines the rate at which scans are processed.
• transform_publish_period: The period for publishing transform data. 0.02 means data is published every 0.02 seconds. If set to 0, it never publishes odometry.
• map_update_interval: The interval at which the map is updated, in seconds.
• resolution: The resolution of the map, in meters.
• max_laser_range: Maximum range for the laser scanner, used in rasterizing images.
• minimum_time_interval: The minimum time interval between scans.
• transform_timeout: Timeout for transform operations.
• tf_buffer_duration: The duration for which transform data is buffered.
• stack_size_to_use: Sets the stack size for the program, important for serializing large maps.
• enable_interactive_mode: Enables or disables interactive mode, allowing for real-time user interaction.

4. General Parameters:

• use_scan_matching: When set to true, the system uses scan matching to align scans with the map.
• use_scan_barycenter: If true, it uses the barycenter (center of mass) of the scan for calculations, which can help in more accurate scan matching.
• minimum_travel_distance: The minimum distance the robot must travel before processing another scan, specified in meters.
• minimum_travel_heading: The minimum change in heading or orientation required before processing another scan.
• scan_buffer_size: The number of scans to keep in the buffer for processing.
• scan_buffer_maximum_scan_distance: The maximum distance a scan can be from the robot for it to be considered for processing.
• link_match_minimum_response_fine: The minimum response value for fine matching when linking scans, used for determining scan quality.
• link_scan_maximum_distance: The maximum distance between scans for them to be linked.
• loop_search_maximum_distance: The maximum distance for searching loop closures.
• do_loop_closing: If set to false, the system does not perform loop closure, which is crucial for correcting drift in long-term mapping.
• loop_match_minimum_chain_size: The minimum size of a chain of scans for loop closure matching.
• loop_match_maximum_variance_coarse: The maximum variance allowed for coarse loop closure matching.
• loop_match_minimum_response_coarse: The minimum response value for coarse loop closure matching.
• loop_match_minimum_response_fine: The minimum response value for fine loop closure matching.

5. Correlation Parameters - General:

• correlation_search_space_dimension: The size of the search space in correlation-based scan matching, in meters.
• correlation_search_space_resolution: The resolution of the search space in correlation-based scan matching.
• correlation_search_space_smear_deviation: The standard deviation of the smear in the correlation search space.

6. Correlation Parameters - Loop Closure:

• loop_search_space_dimension: The size of the search space for loop closure, in meters.
• loop_search_space_resolution: The resolution of the search space for loop closure.
• loop_search_space_smear_deviation: The standard deviation of the smear in the loop closure search space.

7. Scan Matcher Parameters

• distance_variance_penalty: This parameter penalizes the variance in distance measurements. A higher value would mean that the scan matcher is more sensitive to variations in distances measured by the sensors, impacting how scans are aligned and matched.
• angle_variance_penalty: This parameter penalizes the variance in angle measurements. It controls how strictly the scan matcher treats discrepancies in angular measurements, influencing the rotation adjustments made during scan matching.
• fine_search_angle_offset: This is radians used in fine-tuning the scan matching process. It specifies the angular offset used in the finer, more detailed phase of scan matching, allowing for precise alignment of scans.
• coarse_search_angle_offset: This parameter is used in the initial, coarser phase of scan matching. It defines a broader angular offset, helping in the rough alignment of scans before fine-tuning.
• coarse_angle_resolution: This parameter defines the resolution of the angle during the coarse search phase. It determines the granularity of angular adjustments in the initial phase of scan matching.
• minimum_angle_penalty: This is represents the minimum penalty applied for angular discrepancies during scan matching. It affects how discrepancies in angles are weighted in the scan matching algorithm.
• Minimum Distance Penalty (minimum_distance_penalty): Set at 0.5, this parameter specifies the minimum penalty for distance discrepancies. Like the minimum angle penalty, it influences the weight given to distance discrepancies in scan matching.
• use_response_expansion: When set to true, this enables the use of response expansion in scan matching. Response expansion is a technique used to improve the scan matching process, particularly in environments with less distinct features or in situations where initial estimates of the robot's position are less accurate.