329 lines
17 KiB
Plaintext
329 lines
17 KiB
Plaintext
{
|
|
"cells": [
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"RL policy based on the [SoloParkour: Constrained Reinforcement Learning for Visual Locomotion from Privileged Experience](https://arxiv.org/abs/2409.13678). "
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"# Flat Ground"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"## Test In Simulation"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 1,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"ename": "ImportError",
|
|
"evalue": "this platform is not supported: ('failed to acquire X connection: Can\\'t connect to display \":1\": b\\'Authorization required, but no authorization protocol specified\\\\n\\'', DisplayConnectionError(':1', b'Authorization required, but no authorization protocol specified\\n'))\n\nTry one of the following resolutions:\n\n * Please make sure that you have an X server running, and that the DISPLAY environment variable is set correctly",
|
|
"output_type": "error",
|
|
"traceback": [
|
|
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
|
|
"\u001b[0;31mImportError\u001b[0m Traceback (most recent call last)",
|
|
"Cell \u001b[0;32mIn[1], line 2\u001b[0m\n\u001b[1;32m 1\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mGo2Py\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mrobot\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mfsm\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m FSM\n\u001b[0;32m----> 2\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mGo2Py\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mrobot\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mremote\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m KeyboardRemote\n\u001b[1;32m 3\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mGo2Py\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mrobot\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01msafety\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m SafetyHypervisor\n\u001b[1;32m 4\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mGo2Py\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01msim\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01mmujoco\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m Go2Sim\n",
|
|
"File \u001b[0;32m/home/Go2py/Go2Py/robot/remote.py:2\u001b[0m\n\u001b[1;32m 1\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mthreading\u001b[39;00m\n\u001b[0;32m----> 2\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mpynput\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m keyboard\n\u001b[1;32m 5\u001b[0m \u001b[38;5;28;01mclass\u001b[39;00m \u001b[38;5;21;01mBaseRemote\u001b[39;00m:\n\u001b[1;32m 6\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21m__init__\u001b[39m(\u001b[38;5;28mself\u001b[39m):\n",
|
|
"File \u001b[0;32m/usr/local/lib/python3.10/dist-packages/pynput/__init__.py:40\u001b[0m\n\u001b[1;32m 34\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mlogging\u001b[39;00m\n\u001b[1;32m 35\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m logging\u001b[38;5;241m.\u001b[39mgetLogger(\u001b[38;5;124m'\u001b[39m\u001b[38;5;132;01m{}\u001b[39;00m\u001b[38;5;124m.\u001b[39m\u001b[38;5;132;01m{}\u001b[39;00m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m.\u001b[39mformat(\n\u001b[1;32m 36\u001b[0m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124m.\u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m.\u001b[39mjoin(\u001b[38;5;28mcls\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__module__\u001b[39m\u001b[38;5;241m.\u001b[39msplit(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124m.\u001b[39m\u001b[38;5;124m'\u001b[39m, \u001b[38;5;241m2\u001b[39m)[:\u001b[38;5;241m2\u001b[39m]),\n\u001b[1;32m 37\u001b[0m \u001b[38;5;28mcls\u001b[39m\u001b[38;5;241m.\u001b[39m\u001b[38;5;18m__name__\u001b[39m))\n\u001b[0;32m---> 40\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m keyboard\n\u001b[1;32m 41\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01m.\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m mouse\n",
|
|
"File \u001b[0;32m/usr/local/lib/python3.10/dist-packages/pynput/keyboard/__init__.py:31\u001b[0m\n\u001b[1;32m 26\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mitertools\u001b[39;00m\n\u001b[1;32m 28\u001b[0m \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;21;01mpynput\u001b[39;00m\u001b[38;5;21;01m.\u001b[39;00m\u001b[38;5;21;01m_util\u001b[39;00m \u001b[38;5;28;01mimport\u001b[39;00m backend, Events\n\u001b[0;32m---> 31\u001b[0m backend \u001b[38;5;241m=\u001b[39m \u001b[43mbackend\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;18;43m__name__\u001b[39;49m\u001b[43m)\u001b[49m\n\u001b[1;32m 32\u001b[0m KeyCode \u001b[38;5;241m=\u001b[39m backend\u001b[38;5;241m.\u001b[39mKeyCode\n\u001b[1;32m 33\u001b[0m Key \u001b[38;5;241m=\u001b[39m backend\u001b[38;5;241m.\u001b[39mKey\n",
|
|
"File \u001b[0;32m/usr/local/lib/python3.10/dist-packages/pynput/_util/__init__.py:77\u001b[0m, in \u001b[0;36mbackend\u001b[0;34m(package)\u001b[0m\n\u001b[1;32m 74\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m module \u001b[38;5;129;01min\u001b[39;00m RESOLUTIONS:\n\u001b[1;32m 75\u001b[0m resolutions\u001b[38;5;241m.\u001b[39mappend(RESOLUTIONS[module])\n\u001b[0;32m---> 77\u001b[0m \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mImportError\u001b[39;00m(\u001b[38;5;124m'\u001b[39m\u001b[38;5;124mthis platform is not supported: \u001b[39m\u001b[38;5;132;01m{}\u001b[39;00m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m.\u001b[39mformat(\n\u001b[1;32m 78\u001b[0m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124m; \u001b[39m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m.\u001b[39mjoin(\u001b[38;5;28mstr\u001b[39m(e) \u001b[38;5;28;01mfor\u001b[39;00m e \u001b[38;5;129;01min\u001b[39;00m errors)) \u001b[38;5;241m+\u001b[39m (\u001b[38;5;124m'\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;124m'\u001b[39m\n\u001b[1;32m 79\u001b[0m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124mTry one of the following resolutions:\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;124m'\u001b[39m\n\u001b[1;32m 80\u001b[0m \u001b[38;5;241m+\u001b[39m \u001b[38;5;124m'\u001b[39m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;130;01m\\n\u001b[39;00m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m.\u001b[39mjoin(\n\u001b[1;32m 81\u001b[0m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124m * \u001b[39m\u001b[38;5;132;01m{}\u001b[39;00m\u001b[38;5;124m'\u001b[39m\u001b[38;5;241m.\u001b[39mformat(s)\n\u001b[1;32m 82\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m s \u001b[38;5;129;01min\u001b[39;00m resolutions))\n\u001b[1;32m 83\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m resolutions \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;124m'\u001b[39m\u001b[38;5;124m'\u001b[39m)\n",
|
|
"\u001b[0;31mImportError\u001b[0m: this platform is not supported: ('failed to acquire X connection: Can\\'t connect to display \":1\": b\\'Authorization required, but no authorization protocol specified\\\\n\\'', DisplayConnectionError(':1', b'Authorization required, but no authorization protocol specified\\n'))\n\nTry one of the following resolutions:\n\n * Please make sure that you have an X server running, and that the DISPLAY environment variable is set correctly"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"from Go2Py.robot.fsm import FSM\n",
|
|
"from Go2Py.robot.remote import KeyboardRemote\n",
|
|
"from Go2Py.robot.safety import SafetyHypervisor\n",
|
|
"from Go2Py.sim.mujoco import Go2Sim\n",
|
|
"from Go2Py.control.cat import *\n",
|
|
"import torch"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 2,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"ename": "NameError",
|
|
"evalue": "name 'Go2Sim' is not defined",
|
|
"output_type": "error",
|
|
"traceback": [
|
|
"\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
|
|
"\u001b[0;31mNameError\u001b[0m Traceback (most recent call last)",
|
|
"Cell \u001b[0;32mIn[2], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m robot \u001b[38;5;241m=\u001b[39m \u001b[43mGo2Sim\u001b[49m()\n",
|
|
"\u001b[0;31mNameError\u001b[0m: name 'Go2Sim' is not defined"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"robot = Go2Sim()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 3,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"remote = KeyboardRemote()\n",
|
|
"robot.sitDownReset()\n",
|
|
"safety_hypervisor = SafetyHypervisor(robot)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 4,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"class CaTController:\n",
|
|
" def __init__(self, robot, remote, checkpoint):\n",
|
|
" self.remote = remote\n",
|
|
" self.robot = robot\n",
|
|
" self.policy = Policy(checkpoint)\n",
|
|
" self.command_profile = CommandInterface()\n",
|
|
" self.agent = CaTAgent(self.command_profile, self.robot)\n",
|
|
" self.init()\n",
|
|
" self.hist_data = {}\n",
|
|
"\n",
|
|
" def init(self):\n",
|
|
" self.obs = self.agent.reset()\n",
|
|
" self.policy_info = {}\n",
|
|
" self.command_profile.yaw_vel_cmd = 0.0\n",
|
|
" self.command_profile.x_vel_cmd = 0.0\n",
|
|
" self.command_profile.y_vel_cmd = 0.0\n",
|
|
"\n",
|
|
" def update(self, robot, remote):\n",
|
|
" action = self.policy(self.obs, self.policy_info)\n",
|
|
" self.obs, self.ret, self.done, self.info = self.agent.step(action)\n",
|
|
" for key, value in self.info.items():\n",
|
|
" if key in self.hist_data:\n",
|
|
" self.hist_data[key].append(value)\n",
|
|
" else:\n",
|
|
" self.hist_data[key] = [value]"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 5,
|
|
"metadata": {},
|
|
"outputs": [
|
|
{
|
|
"name": "stderr",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"/home/Go2py/Go2Py/control/cat.py:94: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
|
|
" actor_sd = torch.load(checkpoint_path)\n"
|
|
]
|
|
},
|
|
{
|
|
"name": "stdout",
|
|
"output_type": "stream",
|
|
"text": [
|
|
"p_gains: [20. 20. 20. 20. 20. 20. 20. 20. 20. 20. 20. 20.]\n"
|
|
]
|
|
}
|
|
],
|
|
"source": [
|
|
"from Go2Py import ASSETS_PATH \n",
|
|
"import os\n",
|
|
"checkpoint_path = os.path.join(ASSETS_PATH, 'checkpoints/SoloParkour/go2_flat.pt')\n",
|
|
"controller = CaTController(robot, remote, checkpoint_path)\n",
|
|
"fsm = FSM(robot, remote, safety_hypervisor, user_controller_callback=controller.update)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 6,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"controller.command_profile.x_vel_cmd=-0.3\n",
|
|
"controller.command_profile.y_vel_cmd=0.3\n",
|
|
"controller.command_profile.yaw_vel_cmd = 0.0"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"Pressing `u` on the keyboard will make the robot stand up. This is equivalent to the `L2+A` combo of the Go2 builtin state machine. After the the robot is on its feet, pressing `s` will hand over the control the RL policy. This action is equivalent to the `start` key of the builtin controller. When you want to stop, pressing `u` again will act similarly to the real robot and locks it in standing mode. Finally, pressing `u` again will command the robot to sit down."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": 7,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"fsm.close()"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"## Test on Real Robot (ToDo)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from Go2Py.robot.fsm import FSM\n",
|
|
"from Go2Py.robot.remote import UnitreeRemote\n",
|
|
"from Go2Py.robot.safety import SafetyHypervisor\n",
|
|
"from Go2Py.control.walk_these_ways import *"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"from Go2Py.robot.interface import GO2Real\n",
|
|
"import numpy as np\n",
|
|
"robot = GO2Real(mode='lowlevel')"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"remote = UnitreeRemote(robot)\n",
|
|
"safety_hypervisor = SafetyHypervisor(robot)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"class walkTheseWaysController:\n",
|
|
" def __init__(self, robot, remote, checkpoint):\n",
|
|
" self.remote = remote\n",
|
|
" self.robot = robot\n",
|
|
" self.cfg = loadParameters(checkpoint)\n",
|
|
" self.policy = Policy(checkpoint)\n",
|
|
" self.command_profile = CommandInterface()\n",
|
|
" self.agent = WalkTheseWaysAgent(self.cfg, self.command_profile, self.robot)\n",
|
|
" self.agent = HistoryWrapper(self.agent)\n",
|
|
" self.init()\n",
|
|
"\n",
|
|
" def init(self):\n",
|
|
" self.obs = self.agent.reset()\n",
|
|
" self.policy_info = {}\n",
|
|
" self.command_profile.yaw_vel_cmd = 0.0\n",
|
|
" self.command_profile.x_vel_cmd = 0.0\n",
|
|
" self.command_profile.y_vel_cmd = 0.0\n",
|
|
" self.command_profile.stance_width_cmd=0.25\n",
|
|
" self.command_profile.footswing_height_cmd=0.08\n",
|
|
" self.command_profile.step_frequency_cmd = 3.0\n",
|
|
" self.command_profile.bodyHeight = 0.00\n",
|
|
"\n",
|
|
" def update(self, robot, remote):\n",
|
|
" action = self.policy(self.obs, self.policy_info)\n",
|
|
" self.obs, self.ret, self.done, self.info = self.agent.step(action)\n",
|
|
" vy = -robot.getRemoteState().lx\n",
|
|
" vx = robot.getRemoteState().ly\n",
|
|
" omega = -robot.getRemoteState().rx*2.2\n",
|
|
" self.command_profile.x_vel_cmd = vx*1.5\n",
|
|
" self.command_profile.y_vel_cmd = vy*1.5\n",
|
|
" self.command_profile.yaw_vel_cmd = omega"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"checkpoint = \"../Go2Py/assets/checkpoints/walk_these_ways/\"\n",
|
|
"controller = walkTheseWaysController(robot, remote, checkpoint)\n",
|
|
"safety_hypervisor = SafetyHypervisor(robot)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"controller.command_profile.pitch_cmd=0.0\n",
|
|
"controller.command_profile.body_height_cmd=0.0\n",
|
|
"controller.command_profile.footswing_height_cmd=0.04\n",
|
|
"controller.command_profile.roll_cmd=0.0\n",
|
|
"controller.command_profile.stance_width_cmd=0.2\n",
|
|
"controller.command_profile.x_vel_cmd=-0.2\n",
|
|
"controller.command_profile.y_vel_cmd=0.01\n",
|
|
"controller.command_profile.setGaitType(\"trotting\")"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"fsm = FSM(robot, remote, safety_hypervisor, user_controller_callback=controller.update)"
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "markdown",
|
|
"metadata": {},
|
|
"source": [
|
|
"Pressing `L2+A` to command the robot to stand up. After the the robot is on its feet, pressing `start` will hand over the control the RL policy. When you want to stop, pressing `L2+A` again will act similarly to the factory controller and locks the robot in standing mode. Finally, pressing `L2+A` again will command the robot to sit down."
|
|
]
|
|
},
|
|
{
|
|
"cell_type": "code",
|
|
"execution_count": null,
|
|
"metadata": {},
|
|
"outputs": [],
|
|
"source": [
|
|
"fsm.close()"
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"kernelspec": {
|
|
"display_name": "b1-env",
|
|
"language": "python",
|
|
"name": "python3"
|
|
},
|
|
"language_info": {
|
|
"codemirror_mode": {
|
|
"name": "ipython",
|
|
"version": 3
|
|
},
|
|
"file_extension": ".py",
|
|
"mimetype": "text/x-python",
|
|
"name": "python",
|
|
"nbconvert_exporter": "python",
|
|
"pygments_lexer": "ipython3",
|
|
"version": "3.10.12"
|
|
}
|
|
},
|
|
"nbformat": 4,
|
|
"nbformat_minor": 2
|
|
}
|