selected-models/phase_d_reward_config.json
| 1 | { |
| 2 | "failed_landing_penalty": 800.0, |
| 3 | "flip_completion_bonus": 500.0, |
| 4 | "flip_landing_bonus": 2000.0, |
| 5 | "in_zone_crash_penalty": 1200.0, |
| 6 | "landing_without_flip_penalty": 300.0, |
| 7 | "landing_zone_half_width": 0.2, |
| 8 | "near_ground_height": 0.6, |
| 9 | "near_ground_overspeed_weight": 120.0, |
| 10 | "no_flip_terminal_penalty": 300.0, |
| 11 | "outside_zone_landing_penalty": 500.0, |
| 12 | "post_flip_angle_weight": 70.0, |
| 13 | "post_flip_angular_speed_weight": 40.0, |
| 14 | "post_flip_center_weight": 80.0, |
| 15 | "post_flip_horizontal_deadband": 0.08, |
| 16 | "post_flip_horizontal_speed_weight": 50.0, |
| 17 | "post_flip_leg_contact_weight": 30.0, |
| 18 | "post_flip_max_target_vx": 0.35, |
| 19 | "post_flip_original_reward_weight": 3.0, |
| 20 | "post_flip_shaping_clip": 30.0, |
| 21 | "post_flip_shaping_gamma": 0.999, |
| 22 | "post_flip_shaping_weight": 2.0, |
| 23 | "post_flip_target_vx_gain": 0.5, |
| 24 | "post_flip_target_vy_high": -0.45, |
| 25 | "post_flip_target_vy_near_ground": -0.12, |
| 26 | "post_flip_vertical_speed_weight": 100.0, |
| 27 | "post_flip_zone_excess_weight": 120.0, |
| 28 | "pre_flip_original_reward_weight": 0.15, |
| 29 | "recovery_angular_velocity_tolerance": 0.5, |
| 30 | "recovery_bonus": 250.0, |
| 31 | "required_rotations": 1.0, |
| 32 | "rotation_direction": 1, |
| 33 | "rotation_progress_bonus": 300.0, |
| 34 | "safe_touchdown_vertical_speed": 0.18, |
| 35 | "upright_tolerance_radians": 0.3 |
| 36 | } |