Install the SENSOR module

This is an optional module that enables the modeling and simulation of sensors for the purpose of simulating robotics and autonomous agents within Chrono.

Read the introduction to modules for a technical background on the modularity of the Chrono project.

Features

The SENSOR module allows users to model and simulate sensors for robotics and autonomous agents.

For more detail, read the Chrono::Sensor section of the reference manual.

Required Dependencies

  • To run applications based on this module, the following are required:
    • NVIDIA GPU, Maxwell or later - capable of running OptiX
  • To build applications based on this module, the following are required:
    • CUDA
    • OptiX - version 7.2 only (will NOT work with 6.X or 7.3)
    • GLFW - version 3.0 or later
    • GLEW - version 1.0 or later
    • OpenGL
    • TensoRT (optional) - version 7.0.0

Building instructions

  1. Repeat the instructions for the full installation, but when you see the CMake window, you must add the following steps:
  2. Set the ENABLE_MODULE_SENSOR as 'on', then press 'Configure' (to refresh the variable list)
  3. Set the OPTIX_INSTALL_DIR variable to the root of the OptiX directory installed on the system (directory that includes include/) and press 'Configure' to refresh the variable list. If a refresh does not correctly set the following variable: OptiX_Include manually set it accordingly (e.g. OPTIX_INSTALL_DIR/include).
  4. Set all the values for GLEW_..., GLFW_... to proper directory or file values if not automatically found by cmake.
  5. Optionally set USE_CUDA_NVRTC to 'on' to enable runtime compilation of the Optix RT Kernels. Press 'Configure' to refresh the variable list. If set to 'off', the RT Kernels will be compiled at runtime. Depending on the system, you may need to set CUDA_ARCH_NAME to the specific target architecture since this will result in RT Kernels being compiled to PTX.
  6. Optionally set USE_TENSOR_RT to 'on' to enable use of TensorRT for augmenting sensor data. Press 'Configure' to refresh the variable list.
    • Set the TENSOR_RT_INSTALL_DIR variable to the root of the TensorRT directory installed on the system (directory that includes lib/, bin/, include/) and press 'Configure to refresh the avriable list'
    • If a refresh does not correctly set the following variables: TENSOR_RT_INCLUDE_PATH,TENSOR_RT_NVINFER,TENSOR_RT_ONNXPARSER, and TENSOR_RT_PARSER, manually set them accordingly with the last three variable pointing directly to their corresponding library files.
  7. Press 'Configure' again, then 'Generate', and proceed as usual in the installation instructions.
  8. NOTE if linking to Chrono::Sensor install from an external project, make sure to set the directory of the install location where the shader code (compiled ptx code or shaders/*.cu files) is located. This should be set at the top of any external code that will use chrono::sensor from an install location.
    //function to set the shader location (include ChOptixUtils.h)
    chrono::sensor::SetSensorShaderDir("path/to/sensor/shaders");
    //if USE_CUDA_NVRTC is enabled, use
    chrono::sensor::SetSensorShaderDir("path/to/install/include/chrono_sensor/optix/shaders/");
    //if USE_CUDA_NVRTC is disabled, use
    chrono::sensor::SetSensorShaderDir("path/to/install/lib/sensor_ptx/");

    How to use it

  • Consult the reference manual.
  • Look at the API section of this module for documentation about classes and functions.
  • Look at the C++ and Python source of demos to learn how to use the functions of this module.

MacOS support

This module is not supported on the Mac.