chibi314
commented
7 years ago Tuning parameters
-
Vision detection h_max, h_min, s_max, s_min, v_max, v_min camera_offset_x : the yaw offset of the cam is dorminant. yawのオフセットが大きい！カメラマウントのyaw軸を調整！ first_detection_depth_thre tree_diameter
-
Laser clustering max_radius min_points
-
Laser shadow filter min_angle, max_angle neighbors window
-
Laser scan range filter lower_threshold upper_threshold
-
Tree database filter_rate tree_margin_radius valid_num
-
Tree tracking(tree_detector.launch) uav_tilt_thre first_tree_pos_margin first_tree_dist_thresh tree_radius_max tree_radius_min tree_scan_angle_thre tree_circle_regulation_thre narrow_searching_radius narrow_angle_diff_min
-
Motion (motion.yaml)
navigation
takeoff_height : 1.6 nav_xy_pos_pgain : 0.9 nav_z_pos_pgain : 0.8 nav_yaw_pgain : 1.2 nav_xy_vel_thresh : 0.5 nav_z_vel_thresh : 1.0 nav_yaw_vel_thresh : 0.6 nav_pos_convergence_thresh : 0.1 nav_yaw_convergence_thresh : 0.15 nav_vel_converge0.8_thresh : 1.0 takeoff_forward_offset : 4.0 # positive value deep_return_dist : -1.2 # negative value turn_radius_offset : -1.0 # negative value
circle motion
circle_radius : 1.1 circle_y_vel : 0.3 circle_motion_height_step : 0.5
obstacle avoidance
cluster_num_min : 10 safe_zone_radius : 1.0 drone_obstacle_ignore_maximum_radius : 0.90 drone_safety_minimum_radius : 0.85 avoid_vel : 0.5
-
Result tree_cut_rate
-
Odom generator lidar_tc lidar_noise_cut_thresh guidance_vel_weight
Task params(challenge.launch) task_kind circle_motion_times target_num
@KahnShi @chibi314 We have to memorize the parameters which we have to tuning during the competition.
試合中、もしうまく行かなった場合にパラメータ調整が発生するが、 どのパラメータをいじるべきか、全部リストアップしましょう。 MBZIRCのtask1では、それをやっていなかったからうまく行かなかった説。