BiguaSim

BiguaSim is a high-fidelity simulator for underwater robotics built on top of Unreal Engine 4, and forked from Holodeck by BYU's PCCL Lab.

If you use BiguaSim for your research, please cite our ICRA publication;

@inproceedings{Potokar22icra,
  author = {E. Potokar and S. Ashford and M. Kaess and J. Mangelson},
  title = {Holo{O}cean: An Underwater Robotics Simulator},
  booktitle = {Proc. IEEE Intl. Conf. on Robotics and Automation, ICRA},
  address = {Philadelphia, PA, USA},
  month = may,
  year = {2022}
}

Features

• 3+ rich worlds with various infrastructure for generating data or testing underwater algorithms
• Complete with common underwater sensors including DVL, IMU, optical camera, various sonar, depth sensor, and more
• Highly and easily configurable sensors and missions
• Multi-agent missions, including optical and acoustic communications
• Novel sonar simulation framework for simulating imaging, profiling, sidescan, and echosounder sonars
• Imaging sonar implementation includes realistic noise modeling for small sim-2-real gap
• Easy installation and simple, OpenAI Gym-like Python interface
• High performance - simulation speeds of up to 2x real time are possible. Performance penalty only for what you need
• Run headless or watch your agents learn
• Linux and Windows support

Installation

pip install biguasim

(requires >= Python 3.6)

See Installation for complete instructions.

• Docker usage: docker run -it \ --name biguasim \ --privileged \ --ipc=host \ --gpus all \ --runtime=nvidia \ -v /tmp/.X11-unix:/tmp/.X11-unix \ -v $HOME/path/to/biguasim/content:/home/biguasim/workspace \ -e DISPLAY=$DISPLAY \ -e NVIDIA_VISIBLE_DEVICES=all \ -e NVIDIA_DRIVER_CAPABILITIES=compute,graphics,utility \ -e QT_X11_NO_MITSHM=1 \ mgmateus/hydrone:biguasim-base

  To use your local worlds add to docker command:

  -v $HOME/path/to/your/worlds:/home/biguasim/.local/share/biguasim/1.0.0/worlds \

  su biguasim

  • password sudo : biguasim

  cd workspace/biguasim

  sudo pip install .

Documentation

• Quickstart
• Changelog
• Examples
• Agents
• Sensors
• Available Packages and Worlds
• Docs

Usage Overview

BiguaSim's interface is similar to OpenAI's Gym.

We try and provide a batteries included approach to let you jump right into using BiguaSim, with minimal fiddling required.

To demonstrate, here is a quick example using the DefaultWorlds package:

import biguasim

# Load the environment. This environment contains a hovering AUV in a pier
env = biguasim.make("PierHarbor-Hovering")

# You must call `.reset()` on a newly created environment before ticking/stepping it
env.reset()                         

# The AUV takes commands for each thruster
command = [0, 0, 0, 0, 10, 10, 10, 10]   

for i in range(30):
    state = env.step(command)  

• state: dict of sensor name to the sensor's value (nparray).

If you want to access the data of a specific sensor, import sensors and retrieving the correct value from the state dictionary:

print(state["DVLSensor"])

Multi Agent-Environments

BiguaSim supports multi-agent environments.

Calls to step only provide an action for the main agent, and then tick the simulation.

act provides a persistent action for a specific agent, and does not tick the simulation. After an action has been provided, tick will advance the simulation forward. The action is persisted until another call to act provides a different action.

import biguasim
import numpy as np

env = biguasim.make("Dam-Hovering")
env.reset()

# Provide an action for each agent
env.act('auv0', np.array([0, 0, 0, 0, 10, 10, 10, 10]))

# Advance the simulation
for i in range(300):
  # The action provided above is repeated
  states = env.tick()

You can access the sensor states as before:

dvl = states["auv0"]["DVLSensor"]
location = states["auv0"]["DepthSensor"]

(auv0 comes from the scenario configuration file)

Running BiguaSim Headless

BiguaSim can run headless with GPU accelerated rendering. See Using BiguaSim Headless