Table 1 Range of curriculum parameters.
From: Learning agility and adaptive legged locomotion via curricular hindsight reinforcement learning
| Â | Start | End | Unit |
|---|---|---|---|
Reward factor | Â | Â | |
 Angle velocity | 0 | 0.21 | – |
 Joint torque penalties | 0 | 0.018 | – |
 Base acceleration penalties | 0 | 0.1 | – |
 Energy | 0 | 0.012 | – |
 Foot slip | 0 | 0.3 | – |
Domain randomization | Â | Â | |
 Contact friction | [1.25, 1.25] | [0.4, 1.25] | – |
 Payload mass | \([-1.0, -1.0]\) | \([-1.0, 5.0]\) | kg |
 External disturbance force | [0, 0] | [0, 15] | N |
 External disturbance torque | [0, 0] | [0, 5] | Nm |
 Gravity direction shift | [0, 0] | [0, 15] | degree |
Terrians | Â | Â | |
 Height fields range | [0, 0] | \([-2.5, 2.5]\) | cm |
Commands | Â | Â | |
 Linear velocities | [0, 0.8] | [0, 3.5] | m/s |
 Direction of velocities | \([-0.5, 0.5]\) | \([-\pi , \pi ]\) | rad |
