Swimmer

This environment is part of the Mujoco environments which contains general information about the environment.

Action Space	Box(-1.0, 1.0, (2,), float32)
Observation Space	Box(-inf, inf, (8,), float64)
import	gymnasium.make("Swimmer-v5")

Description

This environment corresponds to the Swimmer environment described in Rémi Coulom’s PhD thesis “Reinforcement Learning Using Neural Networks, with Applications to Motor Control”. The environment aims to increase the number of independent state and control variables compared to classical control environments. The swimmers consist of three or more segments (’links’) and one less articulation joints (’rotors’) - one rotor joint connects exactly two links to form a linear chain. The swimmer is suspended in a two-dimensional pool and always starts in the same position (subject to some deviation drawn from a uniform distribution), and the goal is to move as fast as possible towards the right by applying torque to the rotors and using fluid friction.

Notes

The problem parameters are: Problem parameters:

• n: number of body parts

• m_i: mass of part i (i ∈ {1…n})

• l_i: length of part i (i ∈ {1…n})

• k: viscous-friction coefficient

While the default environment has n = 3, l_i = 0.1, and k = 0.1. It is possible to pass a custom MuJoCo XML file during construction to increase the number of links, or to tweak any of the parameters.

Action Space

The action space is a Box(-1, 1, (2,), float32). An action represents the torques applied between links

Num	Action	Control Min	Control Max	Name (in corresponding XML file)	Joint	Type (Unit)
0	Torque applied on the first rotor	-1	1	motor1_rot	hinge	torque (N m)
1	Torque applied on the second rotor	-1	1	motor2_rot	hinge	torque (N m)

Observation Space

The observation space consists of the following parts (in order):

• qpos (3 elements by default): Position values of the robot’s body parts.

• qvel (5 elements): The velocities of these individual body parts (their derivatives).

By default, the observation does not include the x- and y-coordinates of the front tip. These can be included by passing exclude_current_positions_from_observation=False during construction. In this case, the observation space will be a Box(-Inf, Inf, (10,), float64), where the first two observations are the x- and y-coordinates of the front tip. Regardless of whether exclude_current_positions_from_observation is set to True or False, the x- and y-coordinates are returned in info with the keys "x_position" and "y_position", respectively.

By default, however, the observation space is a Box(-Inf, Inf, (8,), float64) where the elements are as follows:

Num	Observation	Min	Max	Name (in corresponding XML file)	Joint	Type (Unit)
0	angle of the front tip	-Inf	Inf	free_body_rot	hinge	angle (rad)
1	angle of the first rotor	-Inf	Inf	motor1_rot	hinge	angle (rad)
2	angle of the second rotor	-Inf	Inf	motor2_rot	hinge	angle (rad)
3	velocity of the tip along the x-axis	-Inf	Inf	slider1	slide	velocity (m/s)
4	velocity of the tip along the y-axis	-Inf	Inf	slider2	slide	velocity (m/s)
5	angular velocity of front tip	-Inf	Inf	free_body_rot	hinge	angular velocity (rad/s)
6	angular velocity of first rotor	-Inf	Inf	motor1_rot	hinge	angular velocity (rad/s)
7	angular velocity of second rotor	-Inf	Inf	motor2_rot	hinge	angular velocity (rad/s)
excluded	position of the tip along the x-axis	-Inf	Inf	slider1	slide	position (m)
excluded	position of the tip along the y-axis	-Inf	Inf	slider2	slide	position (m)

Rewards

The total reward is: reward = forward_reward - ctrl_cost.

• forward_reward: A reward for moving forward, this reward would be positive if the Swimmer moves forward (in the positive \(x\) direction / in the right direction). \(w_{forward} \times \frac{dx}{dt}\), where \(dx\) is the displacement of the (front) “tip” (\(x_{after-action} - x_{before-action}\)), \(dt\) is the time between actions, which depends on the frame_skip parameter (default is 4), and frametime which is \(0.01\) - so the default is \(dt = 4 \times 0.01 = 0.04\), \(w_{forward}\) is the forward_reward_weight (default is \(1\)).

• ctrl_cost: A negative reward to penalize the Swimmer for taking actions that are too large. \(w_{control} \times \|action\|_2^2\), where \(w_{control}\) is ctrl_cost_weight (default is \(10^{-4}\)).

info contains the individual reward terms.

Starting State

The initial position state is \(\mathcal{U}_{[-reset\_noise\_scale \times I_{5}, reset\_noise\_scale \times I_{5}]}\). The initial velocity state is \(\mathcal{U}_{[-reset\_noise\_scale \times I_{5}, reset\_noise\_scale \times I_{5}]}\).

where \(\mathcal{U}\) is the multivariate uniform continuous distribution.

Episode End

Termination

The Swimmer never terminates.

Truncation

The default duration of an episode is 1000 timesteps.

Arguments

Swimmer provides a range of parameters to modify the observation space, reward function, initial state, and termination condition. These parameters can be applied during gymnasium.make in the following way:

import gymnasium as gym
env = gym.make('Swimmer-v5', xml_file=...)

Parameter	Type	Default	Description
xml_file	str	"swimmer.xml"	Path to a MuJoCo model
forward_reward_weight	float	1	Weight for forward_reward term (see Rewards section)
ctrl_cost_weight	float	1e-4	Weight for ctrl_cost term (see Rewards section)
reset_noise_scale	float	0.1	Scale of random perturbations of initial position and velocity (see Starting State section)
exclude_current_positions_from_observation	bool	True	Whether or not to omit the x- and y-coordinates from observations. Excluding the position can serve as an inductive bias to induce position-agnostic behavior in policies (see Observation Space section)

Version History

• v5:

  • Minimum mujoco version is now 2.3.3.

  • Added support for fully custom/third party mujoco models using the xml_file argument (previously only a few changes could be made to the existing models).

  • Added default_camera_config argument, a dictionary for setting the mj_camera properties, mainly useful for custom environments.

  • Added env.observation_structure, a dictionary for specifying the observation space compose (e.g. qpos, qvel), useful for building tooling and wrappers for the MuJoCo environments.

  • Return a non-empty info with reset(), previously an empty dictionary was returned, the new keys are the same state information as step().

  • Added frame_skip argument, used to configure the dt (duration of step()), default varies by environment check environment documentation pages.

  • Restored the xml_file argument (was removed in v4).

  • Added forward_reward_weight, ctrl_cost_weight, to configure the reward function (defaults are effectively the same as in v4).

  • Added reset_noise_scale argument to set the range of initial states.

  • Added exclude_current_positions_from_observation argument.

  • Replaced info["reward_fwd"] and info["forward_reward"] with info["reward_forward"] to be consistent with the other environments.

• v4: All MuJoCo environments now use the MuJoCo bindings in mujoco >= 2.1.3.

• v3: Support for gymnasium.make kwargs such as xml_file, ctrl_cost_weight, reset_noise_scale, etc. rgb rendering comes from tracking camera (so agent does not run away from screen).

• v2: All continuous control environments now use mujoco-py >= 1.50.

• v1: max_time_steps raised to 1000 for robot based tasks. Added reward_threshold to environments.

• v0: Initial versions release.