MSD Pose Estimation

A robot localization stack developed by ITB de Labo for Nakayama Iron Works's MSD700 product to fuse some localization sensors using robot_localization package from ROS.

Installation

Clone this repo to your Catkin workspace's src folder.
Install the dependencies. Navigate to install_requires folder and run noetic_dep.sh.

cd ~/catkin_ws/src/msd_pose_estimation/install_requires
sudo chmod +x noetic_dep.sh
./noetic_dep.sh

Compile the stack.

cd ~/catkin_ws
catkin_make

Overview

Launch file

This stack contains several launch files as follows:

imu_filter.launch : this launch file corrects the IMU measurement by accounting for the earth's magnetic field and the robot's tilt [1]. This uses the "imu_filter_madgwick" node from the "imu_tools" package and the "hardware_state.py" node.

robot_localization.launch : this launches "imu_filter.launch" and "ekf_localization_node" from "robot_localization" package [2].

Nodes

This stack contains several custom nodes as follows:

hardware_state.py : this node subscribes to the "hardware_state" topic that consists of the ros_msd700_msgs::HardwareState type of messages as its main data to process. Please refer to this link for the message details. In the end, this node will publish the "wheel/odom" (nav_msgs::Odometry) topic and the "imu/data" (sensor_msgs::Imu) topic that will be supplied to the robot_localization node.
quat_to_eul.py : this node can be set to subscribe to the arbitrary topics that contain orientation data in quaternion such as nav_msgs::Odometry or sensor_msgs::Imu, then this node will convert those quaternions orientation into Euler angles.

Visualization

This stack provides an R-viz config to visualize both "wheel/odom" and "odometry/filtered" topics by simply running: