CyberRT使用笔记

本文是一篇关于CyberRT使用体验的笔记。首先，简要介绍了CyberRT的基本概念、特点及其在自动驾驶领域的应用。接着，详细描述了作者在使用CyberRT过程中的实际操作步骤，包括环境搭建、模块配置、消息通信等关键环节。此外，文章还分享了一些使用技巧、遇到的问题及解决方案，并对CyberRT的性能和稳定性进行了评价。最后，总结了使用CyberRT的体会和感悟，展望了其在自动驾驶领域的未来发展前景。 

1. 头文件

#include <cyber/cyber.h>

#include <cyber/record/record_reader.h>
#include <cyber/record/record_viewer.h>

2. 预定义

using apollo::cyber::Writer;
using apollo::cyber::Reader;
using apollo::cyber::Time;
using apollo::cyber::Node;

using apollo::drivers::Gps;
using apollo::drivers::Image;
using apollo::drivers::Imu;
using apollo::drivers::PCLPointCloud2Array;

using apollo::cyber::record::RecordReader;
using apollo::cyber::record::RecordViewer;

// reader
using GpsReaderPtr = std::shared_ptr<Reader<Gps>>;
using ImageReaderPtr = std::shared_ptr<Reader<Image>>;
using ImuReaderPtr = std::shared_ptr<Reader<Imu>>;
using PointCloud2ArrayReaderPtr = std::shared_ptr<Reader<PCLPointCloud2Array>>;

// writer
using PointCloud2ArrayWriterPtr = std::shared_ptr<Writer<PCLPointCloud2Array>>;

3. 在线或者离线播包读取数据

3.1. 类定义

class Scheduler {
 public:
  Scheduler();

  ~Scheduler();

  void initialize(const std::vector<std::string>& params);

  void run();

 private:
  void imuHandler(const std::shared_ptr<Imu>& msg);

  void gpsHandler(const std::shared_ptr<Gps>& msg);

  void cloudHandler(const std::shared_ptr<PCLPointCloud2Array>& msg);

 private:
  std::unique_ptr<Node> node_;

  GpsReaderPtr gps_reader_;

  PointCloud2ArrayReaderPtr cloud_reader_;
};

3.2. 初始化

void Scheduler::initialize() {
  gps_reader_ = node_->CreateReader<Gps>("gps_topic", 5, 
      [this](const std::shared_ptr<Gps>& msg) {
        this->gpsHandler(msg);
      });

  cloud_reader_ = node_->CreateReader<PCLPointCloud2Array>("cloud_topic", 5,
      [this](const std::shared_ptr<PCLPointCloud2Array>& msg) {
        this->cloudHandler(msg);
      });

　imu_reader_ = node_->CreateReader<Imu>("imu_topic", 10,
      [this](const std::shared_ptr<Imu>& msg) {
        this->imuHandler(msg);
      });
}

3.3. 回调函数

void Scheduler::gpsHandler(const std::shared_ptr<Gps>& msg) {
  double x = pose_msg->pos().x();
  double y = pose_msg->pos().y();
  double z = pose_msg->pos().z();
}

void Scheduler::cloudHandler(const std::shared_ptr<PCLPointCloud2Array>& msg) {
  pcl::PointCloud<pcl::PointXYZI>::Ptr cloud(new pcl::PointCloud<pcl::PointXYZI>);
  adlab::common::util::fromProto<pcl::PointXYZI>(*(msg->mutable_clouds(0)), cloud->get()));
}

void Scheduler::imuHandler(const std::shared_ptr<Imu>& msg) {
  double time = static_cast<double>(msg->utime()) * 1e-6;

  // acceleration
  const auto& acc = msg->linear_accel();
  Eigen::Vector3d acceleration(acc.x(), acc.y(), acc.z());

  // angular velocity
  const auto& omega = msg->rotation_rate();
  Eigen::Vector3d angular_velocity(omega.x(), omega.y(), omega.z());
}

3.4. 主函数

#include <stdlib.h>
#include <stdio.h>

DEFINE_string(configs, "configs.yaml", "");

int main(int argc, char** argv) {

  google::ParseCommandLineFlags(&argc, &argv, true);

  std::vector<std::string> arguments;
  arguments.push_back(FLAGS_configs);

  adlab::cyber::Init(argv[0]);
  google::SetLogDestination(google::INFO, "");
  ros::init(argc, argv, "node");

  Scheduler schedular;
  schedular.initialize(arguments);

  adlab::cyber::Rate rate(10);
  while (adlab::cyber::OK()) {
    schedular.run();
    rate.Sleep();
  }

  adlab::cyber::WaitForShutdown();
  return 0;
}

4. 从包中直接读取数据

std::shared_ptr<RecordReader> record_reader_ptr = std::make_shared<RecordReader>(file_path);
std::set<std::string> channel_list = record_reader_ptr->GetChannelList();
std::shared_ptr<RecordViewer> recorder_viewer = std::make_shared<RecordViewer>(record_reader_ptr);
for (const auto& msg : *recorder_viewer) {
    if (msg.channel_name == channel_name_) {
        double time = msg.time;
        Gps gps_msg;
        gps_msg.ParseFromString(msg.content);
    }
}

5. 常用命令

5.1. 播放记录

cyber_record play -f data

5.2. 显示记录信息

cyber_record info data