- C++11 or later
- CMake (3.10 or later)
- Ceres Solver (tested with 1.13.0)
- Eigen (tested with 3.3.4)
- Protocol Buffers (tested with 3.6.1.3)
Clone the repository:
git clone [repository url]
Link Eigen libraries to proper directory (if needed). For example:
sudo ln -s /usr/include/eigen3/Eigen /usr/include/Eigen
sudo ln -s /usr/include/eigen3/unsupported/ /usr/include/unsupported
Compile protobufs for employed protobuf version. From cloned directory, execute:
protoc pose.proto --cpp_out=$PWD
protoc measurement.proto --cpp_out=$PWD
CMake is employed for building. From the cloned directory, execute:
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ../
make -j4
Before the first usage, link the params.json file to the build directory. Execute from the build directory:
ln -s ../params.json params.json
To run the non-linear least-squares pose estimator + Kalman filter, execute from the build directory:
./pose_estimator_main
This executable expects measurements (packaged as a protocol buffer of the form specified in measurement.proto) over a named pipe (the path for which is specified by the pipe_path_input field of params.json) at each time-step.
As an example, measurements can be simulated and published over the input named pipe by executing the following from the build directory:
./simulate_measurements
The pose_estimator_main executable has two forms of outputs: 1) publishing the solved-for pose (packaged as a protocol buffer of the form specified in pose.proto) over a named pipe (the path for which is specified by the pipe_path_output field of params.json) at each time-step, and 2) logging data to csv files on disk. The options for enabling and configuring both forms of outputs are located in params.json.
As an example, poses published by pose_estimator_main over the output named pipe can be read and outputted to console by executing the following from the build directory:
./pose_reader_example
To analyse data logged to csv files on disk, some MATLAB scripts are provided. From the scripts directory, run post_analysis_main.m when true pose data is unavailable, or run post_analysis_main_with_truth.m when true pose data is available and is logged to csv file.