1. imu的移植
同之前的类似我们,
- 新增sensor_msg的头文件
#include <sensor_msgs/msg/imu.hpp>
#iinclude <sensor_msgs/msg/magnetic_field.hpp>
-
修改定义imu和mag的pub定义以及msg定义
ros::Publisher
---->rclcpp::Publisher<sensor_msgs::msg::Imu>::SharedPtr
ros::Publisher
---->rclcpp::Publisher<sensor_msgs::msg::MagneticField>::SharedPtr
-
添加依赖
在CMakeLists.txt
新增sensor_msg
的依赖
...
find_package(sensor_msgs REQUIRED)
...
ament_target_dependencies(pibot_driver
rclcpp
...
sensor_msgs)
...
- launch文件新增相应的配置参数
parameters=[
...
{"imu/accelerometer_bias": [0.005436, 0.014684, -0.395418]},
{"imu/gyroscope_bias": [0.005436, 0.014684, -0.395418]},
{"imu/use_accelerometer": True},
{"imu/use_gyroscope": True},
{"imu/use_magnetometer": True},
{"imu/perform_calibration": True},
]
这里有与之前稍有不同的是accelerometer_bias
和gyroscope_bias
其接受的是x y z的三个值
ROS1中这样定义
<rosparam if="$(arg use_imu)">
imu/accelerometer_bias: {x: 0.005436, y: 0.014684, z: -0.395418}
imu/gyroscope_bias: {x: -0.035592, y: 0.080670, z: 0.001216}
</rosparam>`
ROS2的参数还不支持map格式,故稍作修改修改为list(c++中是vector)
- 实现pub,分别发布
imu
和mag
void BaseDriver::update_imu() {
...
auto now = this->get_clock()->now();
if (use_accelerometer_ || use_gyroscope_) {
imu_msg_.header.stamp = now;
imu_msg_.angular_velocity.x = DataHolder::get()->imu_data[3] - gyroscope_bias_[0];
imu_msg_.angular_velocity.y = DataHolder::get()->imu_data[4] - gyroscope_bias_[1];
imu_msg_.angular_velocity.z = DataHolder::get()->imu_data[5] - gyroscope_bias_[2];
imu_msg_.linear_acceleration.x = DataHolder::get()->imu_data[0] - acceleration_bias_[0];
imu_msg_.linear_acceleration.y = DataHolder::get()->imu_data[1] - acceleration_bias_[1];
imu_msg_.linear_acceleration.z = DataHolder::get()->imu_data[2] - acceleration_bias_[2];
imu_pub_->publish(imu_msg_);
}
if (use_magnetometer_) {
if (use_mag_msg_) {
mag_msg_.header.stamp = now;
mag_msg_.magnetic_field.x = (DataHolder::get()->imu_data[6] * MILIGAUSS_TO_TESLA_SCALE - (mag_x_max_ - mag_x_min_) / 2 - mag_x_min_);
mag_msg_.magnetic_field.y = (DataHolder::get()->imu_data[7] * MILIGAUSS_TO_TESLA_SCALE - (mag_y_max_ - mag_y_min_) / 2 - mag_y_min_);
mag_msg_.magnetic_field.z = (DataHolder::get()->imu_data[8] * MILIGAUSS_TO_TESLA_SCALE - (mag_z_max_ - mag_z_min_) / 2 - mag_z_min_);
mag_pub_->publish(mag_msg_);
}
}
...
}
2. 校准的服务移植
ROS1中定义service
来实现对imu
校准的服务,ROS2中同样我们实现一个service
类型我们使用同样的Empty
- 添加头文件以及相应的依赖
#include <std_srvs/srv/empty.hpp>
...
find_package(std_srvs REQUIRED)
...
ament_target_dependencies(pibot_driver
...
std_srvs)
- 注册服务,传入相应的回调
advertiseService
--->create_service
参考https://gitee.com/pibot/ros2_tutorials/tree/master/src/cpp_srvcli/src
create_service<std_srvs::srv::Empty>("imu/calibrate_imu",std::bind(&BaseDriver::calibrateCallback, this, std::placeholders::_1, std::placeholders::_2, std::placeholders::_3));
- 回调实现,回调中置相应的变量
void BaseDriver::calibrateCallback(const std::shared_ptr<rmw_request_id_t> __attribute__((unused)) request_header,
const std::shared_ptr<std_srvs::srv::Empty::Request> __attribute__((unused)) request,
std::shared_ptr<std_srvs::srv::Empty::Response> __attribute__((unused)) response) {
RCLCPP_INFO(this->get_logger(), "Calibrating accelerometer and gyroscope, make sure robot is stationary and level.");
perform_calibration_ = true;
}
3. 测试
ros2 launch pibot_bringup bringup_launch.py
➜ ros2 topic list
/cmd_vel
/imu/data_raw
/imu/mag
/odom
/parameter_events
/rosout
/tf
➜ ros2 topic info /imu/data_raw
Type: sensor_msgs/msg/Imu
Publisher count: 1
Subscription count: 0
➜ ros2 topic info /imu/mag
Type: sensor_msgs/msg/MagneticField
Publisher count: 1
Subscription count: 0
➜ ros2 service list
/imu/calibrate_imu
/pibot_driver/describe_parameters
➜ ros2 service type /imu/calibrate_imu
std_srvs/srv/Empty
调用服务,类似ROS1中的rosserivce call
,ROS2中使用ros2 service call
➜ ros2 service call /imu/calibrate_imu std_srvs/srv/Empty {}
requester: making request: std_srvs.srv.Empty_Request()
response:
std_srvs.srv.Empty_Response()
本文代码https://gitee.com/pibot/pibot_bringup/tree/5e7e0c961a35f2704bb92a1f69fb4af3640d4093
网友评论