How to build your own environment in Robosuite

How to build your own environment in Robosuite

Overview

Robosuite is a highly renowned simulation environment for embodied AI. Numerous projects have been developed based on it, such as Robomimic and MimicGen. Recently, I have been delving deeper into Robosuite because I find it well-suited for my project. Prior to exploring Robosuite, I had also experimented with other simulators and environments, like Sapien and OmniGibson. Each time I start to familiarize myself with a new simulator, the process can be quite challenging and frustrating. Therefore, I am writing this blog to document some useful findings and tips that I discover along the way

how to build your own environment.

First, check the official document. Apparently, you can find it in this page . But when I do as what this page told, I just can not figure out what I am doing. Therefore, I strongly recommend to take a look at those files in *./robosuite/enviroments/manipulation/*. after that you can find there would be 5 functions there for every task respectively, **_load_models, reward, _check_success, reset_internal, _setup_observables**. Just like what has been told in issue#526:  among these 5 functions, the 2 most important ones are _load_models() and _setup_observables. Taking *./robosuite/enviroments/manipulation/lift.py* as an example:

def _load_model(self):
    """
    Loads an xml model, puts it in self.model
    """
    super()._load_model()

    # Adjust base pose accordingly
    xpos = self.robots[0].robot_model.base_xpos_offset["table"](self.table_full_size[0])
    self.robots[0].robot_model.set_base_xpos(xpos)

    # load model for table top workspace
    mujoco_arena = TableArena(
        table_full_size=self.table_full_size,
        table_friction=self.table_friction,
        table_offset=self.table_offset,
    )

    # Arena always gets set to zero origin
    mujoco_arena.set_origin([0, 0, 0])

    # initialize objects of interest
    tex_attrib = {
        "type": "cube",
    }
    mat_attrib = {
        "texrepeat": "1 1",
        "specular": "0.4",
        "shininess": "0.1",
    }
    redwood = CustomMaterial(
        texture="WoodRed",
        tex_name="redwood",
        mat_name="redwood_mat",
        tex_attrib=tex_attrib,
        mat_attrib=mat_attrib,
    )
    self.cube = BoxObject(
        name="cube",
        size_min=[0.020, 0.020, 0.020],  # [0.015, 0.015, 0.015],
        size_max=[0.022, 0.022, 0.022],  # [0.018, 0.018, 0.018])
        rgba=[1, 0, 0, 1],
        material=redwood,
    )

    # Create placement initializer
    if self.placement_initializer is not None:
        self.placement_initializer.reset()
        self.placement_initializer.add_objects(self.cube)
    else:
        self.placement_initializer = UniformRandomSampler(
            name="ObjectSampler",
            mujoco_objects=self.cube,
            x_range=[-0.03, 0.03],
            y_range=[-0.03, 0.03],
            rotation=None,
            ensure_object_boundary_in_range=False,
            ensure_valid_placement=True,
            reference_pos=self.table_offset,
            z_offset=0.01,
        )

    # task includes arena, robot, and objects of interest
    self.model = ManipulationTask(
        mujoco_arena=mujoco_arena,
        mujoco_robots=[robot.robot_model for robot in self.robots],
        mujoco_objects=self.cube,
    )

To design a different scene, we should change the robot, arena, objects here. Besides, you have to register the task in *./robosuite/__init__.py* by adding a line: from robosuite.environment.manipulation.yourtask import yourtask after that, you can modify the parameters in *./robosuite/demos* or add the name of your task after the --environment.

Summary

Robosuite is a great simulator, but its official tutorial isn’t very beginner-friendly. I hope my blog can help you if you run into difficulties while creating your own environment. At the very least, it should save you some time. I wish I had more time to delve into the details in the future.