Apollo9.0 PNC源码学习之Control模块(一)

news2024/12/25 8:55:23

0 前言

从planning的角度看control,首先需要了解的就是相关的数据接口,规划出的轨迹(路径+速度)发给Control模块去执行
modules/planning/planning_component/planning_component.cc
planning模块发布轨迹信息

planning_writer_ = node_->CreateWriter<ADCTrajectory>(
      config_.topic_config().planning_trajectory_topic());

modules/control/control_component/control_component.cc
Control模块接受轨迹信息

trajectory_reader_ =
      node_->CreateReader<ADCTrajectory>(planning_reader_config, nullptr);
ACHECK(trajectory_reader_ != nullptr);

ADCTrajectory在modules/common_msgs/planning_msgs/planning.proto定义

message ADCTrajectory {
  optional apollo.common.Header header = 1;

  optional double total_path_length = 2;  // in meters

  optional double total_path_time = 3;    // in seconds

  optional EStop estop = 6;

  optional apollo.planning_internal.Debug debug = 8;

  // is_replan == true mean replan triggered 重规划 
  optional bool is_replan = 9 [default = false];

  // Specify trajectory gear 档位
  optional apollo.canbus.Chassis.GearPosition gear = 10;

  // path data + speed data 路径数据 + 速度数据
  repeated apollo.common.TrajectoryPoint trajectory_point = 12;

  // path point without speed info 路径点
  repeated apollo.common.PathPoint path_point = 13;

  optional apollo.planning.DecisionResult decision = 14;

  optional LatencyStats latency_stats = 15;

  // the routing used for current planning result
  optional apollo.common.Header routing_header = 16;
  enum RightOfWayStatus {
    UNPROTECTED = 0;
    PROTECTED = 1;
  }
  optional RightOfWayStatus right_of_way_status = 17;

  // lane id along current reference line
  repeated apollo.hdmap.Id lane_id = 18;

  // set the engage advice for based on current planning result.
  optional apollo.common.EngageAdvice engage_advice = 19;

  // the region where planning cares most
  message CriticalRegion {
    repeated apollo.common.Polygon region = 1;
  }
  // critical region will be empty when planning is NOT sure which region is
  // critical
  // critical regions may or may not overlap
  optional CriticalRegion critical_region = 20;
  // 轨迹类型(未知、正常、)
  enum TrajectoryType {
    UNKNOWN = 0;       // 未知的轨迹类型,通常用于表示无法确定或识别的情况
    NORMAL = 1;        // 正常的轨迹类型,可能是由标准路径规划算法生成的轨迹
    PATH_FALLBACK = 2; // 路径回退类型,当标准路径规划失败时,可能会使用备用路径规划算法生成轨迹
    SPEED_FALLBACK = 3;// 速度回退类型,当无法满足速度约束条件时,可能会使用备用速度规划算法生成轨迹
    PATH_REUSED = 4;   // 重用路径类型,可能是之前生成的路径的重用或修改版本
    OPEN_SPACE = 5;    // 开放空间类型,通常用于表示在开放环境中的轨迹规划,比如停车或避障等情况
  }
  optional TrajectoryType trajectory_type = 21 [default = UNKNOWN];

  optional string replan_reason = 22;

  // lane id along target reference line
  repeated apollo.hdmap.Id target_lane_id = 23;

  // complete dead end flag
  optional bool car_in_dead_end = 24;

  // output related to RSS
  optional RSSInfo rss_info = 100;
}

1 纵览控制模块

Control模块由control组件包和controller控制器组成,control组件包包含control的整体架构和流程。control根据上游模块输入planning模块的期望轨迹信息,定位模块的当前定位信息,车辆底盘及车身状态信息,通过不同的控制算法计算控制车辆的指令(包含转向、油门、刹车等)输出给canbus模块
在这里插入图片描述

1.1 control_component

control_component是继承于apollo::cyber::TimerComponent的子类,是一个定时触发的组件,通过dag配置可以修改定时器周期。InitProc是入口函数,在初始化函数中,主要实现了ControlTaskAgent的初始化,以及control上游的相关消息的订阅。在Proc执行函数中,分别执行了几步操作:获取订阅消息的当前最新数据–>检查订阅消息输入数据(代码里主要检查了对轨迹线数据是否为空的检查,其它消息数据的检查也可以自行添加)是否完整–>检查订阅消息输入数据时间戳是否在容差范围内(上游消息的数据周期是否超时,如果超时control会有紧急处理)–>更新车身姿态信息–>进行control控制计算(这部分调用ControlTaskAgent的ComputeControlCommand方法,ControlTaskAgent通过配置文件,管理控制器ControlTask的加载和执行顺序,进而完成控制指令的计算)–>输出底盘控制指令

1.2 control_task_base

control_component/controller_task_base/主要包含ControlTaskAgent和ControlTask定义。ControlTaskAgent用来管理ControlTask插件的加载和执行顺序,ControlTask是controller控制器插件的父类,Control/controller控制器插件都继承于ControlTask,目前Apollo已经支持的控制器插件有横向控制器(LatController),纵向控制器(LonController),MPC控制器(MPCController),以及DemoControlTask任务器(DemoControlTask

1.3 controller

Apollo对车辆的控制是将车辆在车体坐标系转换到Frenet坐标系下进行位置跟踪,将车辆跟踪轨迹的运动分解为横向运动和纵向运动,通过对车体的动力学建模,选取合适的状态变量对车辆的跟踪情况进行观测,再通过横向和纵向的控制算法,计算合理的控制指令,达到对轨迹线的跟踪目标
在这里插入图片描述

1.4 文件组织结构及说明
control/
├── control_component/                  // control基础组件
    ├── common                          // 模块全局gflag定义
    ├── conf                            // 模块配置文件,参数文件目录,包含gflags变量的配置,插件启用的配置文件,车辆标定表等通用的配置文件
    ├── controller_task_base/           // control控制器父类组件
    │   ├── common/                     // 数学公式,算法公式,滤波函数,轨迹分析
    │   ├── integration_tests/          // 单元测试文件夹
    │   ├── control_task_agent.cc       // 控制器加载管理器实现文件
    │   ├── control_task_agent.h        // 控制器加载管理器实现文件
    │   └── control_task.h              // 控制器父类实现文件
    ├── dag/                            // 模块启动文件(mainboard)
    ├── docs/                           // 相关模块说明文档
    ├── launch/                         // 模块启动文件(cyber_launch)
    ├── proto/                          // 组件定义的配置文件
    ├── submodules/                     // control子模块
    ├── testdata/                       // 单元测试数据
    ├── tools/                          // 调试工具
    ├── BUILD                           // 构建规则文件
    ├── control_component.cc            // 组件实现的代码文件
    ├── control_component.h             // 组件实现的代码文件
    ├── control_component_test.cc       // 组件单元测试文件
    ├── control.json                    // 打包描述文件
    ├── cyberfile.xml                   // 包管理配置文件
    └── README_cn.md                    // 说明文档
└── controllers/                        // 控制器算法或逻辑任务组件
    ├── demo_control_task               // demo控制器插件包
    │   ├── proto/                      // 控制器的配置定义文件夹
    │   ├── conf/                       // 控制器配置文件夹
    │   ├── BUILD                       // 构建规则文件
    │   ├── cyberfile.xml               // 包管理配置文件
    │   ├── demo_control_task.cc        // demo控制器实现文件
    │   ├── demo_control_task.h         // demo控制器实现文件
    │   └── plugins.xml                 // 插件规则文件
    ├── lat_based_lqr_controller        // LQR横向控制器插件包
    │   ├── proto/                      // 控制器的配置定义文件夹
    │   ├── conf/                       // 控制器配置文件夹
    │   ├── BUILD                       // 构建规则文件
    │   ├── cyberfile.xml               // 包管理配置文件
    │   ├── lat_controller.cc           // LQR横向控制器实现文件
    │   ├── lat_controller.h            // LQR横向控制器实现文件
    │   ├── lat_controller_test.cc      // LQR横向控制器单元测试文件
    │   ├── lateral_controller_test     // 控制器测试数据
    │   └── plugins.xml                 // 插件规则文件
    ├── lon_based_pid_controller        // PID纵向控制器插件包
    │   ├── proto/                      // 控制器的配置定义文件夹
    │   ├── conf/                       // 控制器配置文件夹
    │   ├── BUILD                       // 构建规则文件
    │   ├── cyberfile.xml               // 包管理配置文件
    │   ├── lon_controller.cc           // PID纵向控制器实现文件
    │   ├── lon_controller.h            // PID纵向控制器实现文件
    │   ├── lon_controller_test.cc      // PID纵向控制器单元测试文件
    │   ├── longitudinal_controller_test// 控制器测试数据
    │   └── plugins.xml                 // 插件规则文件
    └── mpc_controller                  // MPC横纵向控制器插件包
        ├── proto/                      // 控制器的配置定义文件夹
        ├── conf/                       // 控制器配置文件夹
        ├── BUILD                       // 构建规则文件
        ├── cyberfile.xml               // 包管理配置文件
        ├── mpc_controller.cc           // MPC控制器实现文件
        ├── mpc_controller.h            // MPC控制器实现文件
        ├── mpc_controller_test.cc      // MPC控制器单元测试文件
        ├── mpc_controller_test_data    // 控制器测试数据
        └── plugins.xml                 // 插件规则文件
1.5 模块输入输出与配置

输入:

Channel名称类型描述
/apollo/planningapollo::planning::ADCTrajectory车辆规划轨迹线信息
/apollo/localization/poseapollo::localization::LocalizationEstimate车辆定位信息
/apollo/canbus/chassisapollo::canbus::Chassis车辆底盘信息
-apollo::common::VehicleState车身姿态信息
/apollo/control/padapollo::control::ControlCommand::PadMessage自动驾驶使能(请求进入自动驾驶)指令

输出:

Channel名称类型描述
/apollo/controlapollo::control::ControlCommand车辆的控制指令,如方向盘、油门、刹车等信息

配置文件:

文件路径类型/结构说明
modules/control/control_component/conf/pipeline.pb.txtapollo::control::ControlPipelineControlComponent的配置文件
modules/control/control_component/conf/control.confcommand line flags命令行参数配置
modules/control/control_component/conf/calibration_table.pb.txtapollo::control::calibration_table车辆纵向标定表配置

Flags:

flagfile类型描述
modules/control/control_component/common/control_gflags.ccflagsControl组件flags变量定义文件
modules/control/control_component/common/control_gflags.hdeclareControl组件flags声明文件

2 控制器组件代码解析

control_component.h

#pragma once

#include <memory>
#include <string>

#include "modules/common_msgs/chassis_msgs/chassis.pb.h"
#include "modules/common_msgs/control_msgs/control_cmd.pb.h"
#include "modules/common_msgs/control_msgs/pad_msg.pb.h"
#include "modules/common_msgs/external_command_msgs/command_status.pb.h"
#include "modules/common_msgs/localization_msgs/localization.pb.h"
#include "modules/common_msgs/planning_msgs/planning.pb.h"
#include "modules/control/control_component/proto/preprocessor.pb.h"

#include "cyber/class_loader/class_loader.h"
#include "cyber/component/timer_component.h"
#include "cyber/time/time.h"
#include "modules/common/monitor_log/monitor_log_buffer.h"
#include "modules/common/util/util.h"
#include "modules/control/control_component/controller_task_base/common/dependency_injector.h"
#include "modules/control/control_component/controller_task_base/control_task_agent.h"
#include "modules/control/control_component/submodules/preprocessor_submodule.h"

/**
 * @namespace apollo::control
 * @brief apollo::control
 */
namespace apollo {
namespace control {

/**
 * @class Control
 *
 * @brief control module main class, it processes localization, chassis, and
 * pad data to compute throttle, brake and steer values.
 */
// 控制模块主类,处理定位、底盘、pad数据为了计算油门、刹车和转向,继承apollo::cyber::TimerComponent,定时触发
class ControlComponent final : public apollo::cyber::TimerComponent {
  friend class ControlTestBase;

 public:
  ControlComponent();
  bool Init() override;

  bool Proc() override;

 private:
  // Upon receiving pad message
  // 接收pad消息
  void OnPad(const std::shared_ptr<PadMessage> &pad);
  // 接收底盘消息
  void OnChassis(const std::shared_ptr<apollo::canbus::Chassis> &chassis);
  // 接受轨迹消息
  void OnPlanning(
      const std::shared_ptr<apollo::planning::ADCTrajectory> &trajectory);
  // 规划命令状态信息
  void OnPlanningCommandStatus(
      const std::shared_ptr<external_command::CommandStatus>
          &planning_command_status);
  // 接收定位消息
  void OnLocalization(
      const std::shared_ptr<apollo::localization::LocalizationEstimate>
          &localization);

  // Upon receiving monitor message
  // 接收检测信息
  void OnMonitor(
      const apollo::common::monitor::MonitorMessage &monitor_message);

  common::Status ProduceControlCommand(ControlCommand *control_command);
  common::Status CheckInput(LocalView *local_view);
  common::Status CheckTimestamp(const LocalView &local_view);
  common::Status CheckPad();
  void ResetAndProduceZeroControlCommand(ControlCommand *control_command);
  void GetVehiclePitchAngle(ControlCommand *control_command);

 private:
  apollo::cyber::Time init_time_;

  localization::LocalizationEstimate latest_localization_;
  canbus::Chassis latest_chassis_;
  planning::ADCTrajectory latest_trajectory_;
  external_command::CommandStatus planning_command_status_;
  PadMessage pad_msg_;
  common::Header latest_replan_trajectory_header_;

  ControlTaskAgent control_task_agent_;

  bool estop_ = false;
  std::string estop_reason_;
  bool pad_received_ = false;

  unsigned int status_lost_ = 0;
  unsigned int status_sanity_check_failed_ = 0;
  unsigned int total_status_lost_ = 0;
  unsigned int total_status_sanity_check_failed_ = 0;

  ControlPipeline control_pipeline_;

  std::mutex mutex_;
  // 订阅者 底盘、pad、定位、轨迹、命令状态
  std::shared_ptr<cyber::Reader<apollo::canbus::Chassis>> chassis_reader_;
  std::shared_ptr<cyber::Reader<PadMessage>> pad_msg_reader_;
  std::shared_ptr<cyber::Reader<apollo::localization::LocalizationEstimate>>
      localization_reader_;
  std::shared_ptr<cyber::Reader<apollo::planning::ADCTrajectory>>
      trajectory_reader_;
  std::shared_ptr<cyber::Reader<apollo::external_command::CommandStatus>>
      planning_command_status_reader_;
  // 发布者 控制命令、使用控制子模块LocalView
  std::shared_ptr<cyber::Writer<ControlCommand>> control_cmd_writer_;
  // when using control submodules
  std::shared_ptr<cyber::Writer<LocalView>> local_view_writer_;
  
  common::monitor::MonitorLogBuffer monitor_logger_buffer_;

  LocalView local_view_;

  std::shared_ptr<DependencyInjector> injector_;

  double previous_steering_command_ = 0.0;
};

CYBER_REGISTER_COMPONENT(ControlComponent)
}  // namespace control
}  // namespace apollo

control_component.cc

#include "modules/control/control_component/control_component.h"

#include "absl/strings/str_cat.h"

#include "cyber/common/file.h"
#include "cyber/common/log.h"
#include "cyber/time/clock.h"
#include "modules/common/adapters/adapter_gflags.h"
#include "modules/common/latency_recorder/latency_recorder.h"
#include "modules/common/vehicle_state/vehicle_state_provider.h"
#include "modules/control/control_component/common/control_gflags.h"

namespace apollo {
namespace control {

using apollo::canbus::Chassis;
using apollo::common::ErrorCode;
using apollo::common::Status;
using apollo::common::VehicleStateProvider;
using apollo::cyber::Clock;
using apollo::localization::LocalizationEstimate;
using apollo::planning::ADCTrajectory;

const double kDoubleEpsilon = 1e-6;

ControlComponent::ControlComponent()
    : monitor_logger_buffer_(common::monitor::MonitorMessageItem::CONTROL) {}

bool ControlComponent::Init() {
  injector_ = std::make_shared<DependencyInjector>();
  init_time_ = Clock::Now();

  AINFO << "Control init, starting ...";

  ACHECK(
      cyber::common::GetProtoFromFile(FLAGS_pipeline_file, &control_pipeline_))
      << "Unable to load control pipeline file: " + FLAGS_pipeline_file;

  AINFO << "ControlTask pipeline config file: " << FLAGS_pipeline_file
        << " is loaded.";

  // initial controller agent when not using control submodules
  ADEBUG << "FLAGS_use_control_submodules: " << FLAGS_use_control_submodules;
  if (!FLAGS_is_control_ut_test_mode) {
    if (!FLAGS_use_control_submodules &&
        !control_task_agent_.Init(injector_, control_pipeline_).ok()) {
      // set controller
      ADEBUG << "original control";
      monitor_logger_buffer_.ERROR(
          "Control init controller failed! Stopping...");
      return false;
    }
  }

  cyber::ReaderConfig chassis_reader_config;
  chassis_reader_config.channel_name = FLAGS_chassis_topic;
  chassis_reader_config.pending_queue_size = FLAGS_chassis_pending_queue_size;
  // 订阅底盘信息
  chassis_reader_ =
      node_->CreateReader<Chassis>(chassis_reader_config, nullptr);
  ACHECK(chassis_reader_ != nullptr);
  
  cyber::ReaderConfig planning_reader_config;
  planning_reader_config.channel_name = FLAGS_planning_trajectory_topic;
  planning_reader_config.pending_queue_size = FLAGS_planning_pending_queue_size;
  // 订阅轨迹信息
  trajectory_reader_ =
      node_->CreateReader<ADCTrajectory>(planning_reader_config, nullptr);
  ACHECK(trajectory_reader_ != nullptr);

  cyber::ReaderConfig planning_command_status_reader_config;
  planning_command_status_reader_config.channel_name =
      FLAGS_planning_command_status;
  planning_command_status_reader_config.pending_queue_size =
      FLAGS_planning_status_msg_pending_queue_size;
  // 订阅规划命令状态
  planning_command_status_reader_ =
      node_->CreateReader<external_command::CommandStatus>(
          planning_command_status_reader_config, nullptr);
  ACHECK(planning_command_status_reader_ != nullptr);

  cyber::ReaderConfig localization_reader_config;
  localization_reader_config.channel_name = FLAGS_localization_topic;
  localization_reader_config.pending_queue_size =
      FLAGS_localization_pending_queue_size;
  // 订阅定位信息
  localization_reader_ = node_->CreateReader<LocalizationEstimate>(
      localization_reader_config, nullptr);
  ACHECK(localization_reader_ != nullptr);

  cyber::ReaderConfig pad_msg_reader_config;
  pad_msg_reader_config.channel_name = FLAGS_pad_topic;
  pad_msg_reader_config.pending_queue_size = FLAGS_pad_msg_pending_queue_size;
  // 订阅pad消息
  pad_msg_reader_ =
      node_->CreateReader<PadMessage>(pad_msg_reader_config, nullptr);
  ACHECK(pad_msg_reader_ != nullptr);
  // 如果使用控制子模块,发布控制话题,否则,发布控制local_view
  if (!FLAGS_use_control_submodules) {
    control_cmd_writer_ =
        node_->CreateWriter<ControlCommand>(FLAGS_control_command_topic);
    ACHECK(control_cmd_writer_ != nullptr);
  } else {
    local_view_writer_ =
        node_->CreateWriter<LocalView>(FLAGS_control_local_view_topic);
    ACHECK(local_view_writer_ != nullptr);
  }

  // set initial vehicle state by cmd
  // need to sleep, because advertised channel is not ready immediately
  // simple test shows a short delay of 80 ms or so
  // 休眠1000ms
  AINFO << "Control resetting vehicle state, sleeping for 1000 ms ...";
  std::this_thread::sleep_for(std::chrono::milliseconds(1000));

  // should init_vehicle first, let car enter work status, then use status msg
  // trigger control
  // 首先初始化vehicle,让车到工作状态,然后使用状态消息
  AINFO << "Control default driving action is "
        << DrivingAction_Name((enum DrivingAction)FLAGS_action);
  pad_msg_.set_action((enum DrivingAction)FLAGS_action);

  return true;
}

void ControlComponent::OnPad(const std::shared_ptr<PadMessage> &pad) {
  // 创建锁,确保在访问 pad_msg_ 期间不会发生数据竞争
  std::lock_guard<std::mutex> lock(mutex_);
  // 将 pad 中的数据复制到 pad_msg_
  pad_msg_.CopyFrom(*pad);
  ADEBUG << "Received Pad Msg:" << pad_msg_.DebugString();
  AERROR_IF(!pad_msg_.has_action()) << "pad message check failed!";
}

void ControlComponent::OnChassis(const std::shared_ptr<Chassis> &chassis) {
  ADEBUG << "Received chassis data: run chassis callback.";
  // 创建锁,确保在访问 latest_chassis_ 期间不会发生数据竞争
  std::lock_guard<std::mutex> lock(mutex_);
  latest_chassis_.CopyFrom(*chassis);
}

void ControlComponent::OnPlanning(
    const std::shared_ptr<ADCTrajectory> &trajectory) {
  ADEBUG << "Received chassis data: run trajectory callback.";
  // 创建锁,确保在访问 latest_trajectory_ 期间不会发生数据竞争
  std::lock_guard<std::mutex> lock(mutex_);
  latest_trajectory_.CopyFrom(*trajectory);
}

void ControlComponent::OnPlanningCommandStatus(
    const std::shared_ptr<external_command::CommandStatus>
        &planning_command_status) {
  ADEBUG << "Received plannning command status data: run planning command "
            "status callback.";
  // 创建锁,确保在访问 planning_command_status_ 期间不会发生数据竞争
  std::lock_guard<std::mutex> lock(mutex_);
  planning_command_status_.CopyFrom(*planning_command_status);
}

void ControlComponent::OnLocalization(
    const std::shared_ptr<LocalizationEstimate> &localization) {
  ADEBUG << "Received control data: run localization message callback.";
  // 创建锁,确保在访问 latest_localization_ 期间不会发生数据竞争
  std::lock_guard<std::mutex> lock(mutex_);
  latest_localization_.CopyFrom(*localization);
}

void ControlComponent::OnMonitor(
    const common::monitor::MonitorMessage &monitor_message) {
  for (const auto &item : monitor_message.item()) {
    if (item.log_level() == common::monitor::MonitorMessageItem::FATAL) {
      // 检测到严重问题,需要立即停止
      estop_ = true;
      return;
    }
  }
}

Status ControlComponent::ProduceControlCommand(
    ControlCommand *control_command) {
  // 检查输入数据
  Status status = CheckInput(&local_view_);
  // check data
  if (!status.ok()) {
    AERROR_EVERY(100) << "Control input data failed: "
                      << status.error_message();
    control_command->mutable_engage_advice()->set_advice(
        apollo::common::EngageAdvice::DISALLOW_ENGAGE);
    control_command->mutable_engage_advice()->set_reason(
        status.error_message());
    estop_ = true;
    estop_reason_ = status.error_message();
  } else {
    estop_ = false;
    // 检查时间戳
    Status status_ts = CheckTimestamp(local_view_);
    if (!status_ts.ok()) {
      AERROR << "Input messages timeout";
      // Clear trajectory data to make control stop if no data received again
      // next cycle.
      // keep the history trajectory for control compute.
      // latest_trajectory_.Clear();
      estop_ = true;
      status = status_ts;
      if (local_view_.chassis().driving_mode() !=
          apollo::canbus::Chassis::COMPLETE_AUTO_DRIVE) {
        control_command->mutable_engage_advice()->set_advice(
            apollo::common::EngageAdvice::DISALLOW_ENGAGE);
        control_command->mutable_engage_advice()->set_reason(
            status.error_message());
      }
    } else {
      control_command->mutable_engage_advice()->set_advice(
          apollo::common::EngageAdvice::READY_TO_ENGAGE);
      estop_ = false;
    }
  }

  // 检查 estop
  estop_ = FLAGS_enable_persistent_estop
               ? estop_ || local_view_.trajectory().estop().is_estop()
               : local_view_.trajectory().estop().is_estop();
  // 如果规划中的 estop 标志为真,则设置 estop_ 为真
  if (local_view_.trajectory().estop().is_estop()) {
    estop_ = true;
    estop_reason_ = "estop from planning : ";
    estop_reason_ += local_view_.trajectory().estop().reason();
  }
  // 如果规划中的轨迹点为空,则设置 estop_ 为真
  if (local_view_.trajectory().trajectory_point().empty()) {
    AWARN_EVERY(100) << "planning has no trajectory point. ";
    estop_ = true;
    estop_reason_ = "estop for empty planning trajectory, planning headers: " +
                    local_view_.trajectory().header().ShortDebugString();
  }
  // 如果启用了 gear_drive 负速度保护功能,并且当前驾驶模式为 gear_drive,并且第一个轨迹点的速度小于 -kEpsilon,则设置 estop_ 为真
  if (FLAGS_enable_gear_drive_negative_speed_protection) {
    const double kEpsilon = 0.001;
    auto first_trajectory_point = local_view_.trajectory().trajectory_point(0);
    if (local_view_.chassis().gear_location() == Chassis::GEAR_DRIVE &&
        first_trajectory_point.v() < -1 * kEpsilon) {
      estop_ = true;
      estop_reason_ = "estop for negative speed when gear_drive";
    }
  }

  if (!estop_) {
    // 如果当前驾驶模式为完全手动驾驶,则重置控制器
    if (local_view_.chassis().driving_mode() == Chassis::COMPLETE_MANUAL) {
      control_task_agent_.Reset();
      AINFO_EVERY(100) << "Reset Controllers in Manual Mode";
    }
    // 设置控制命令的调试信息
    auto debug = control_command->mutable_debug()->mutable_input_debug();
    debug->mutable_localization_header()->CopyFrom(
        local_view_.localization().header());
    debug->mutable_canbus_header()->CopyFrom(local_view_.chassis().header());
    debug->mutable_trajectory_header()->CopyFrom(
        local_view_.trajectory().header());
    // 如果当前规划的轨迹点不为空,则将最新的重新规划轨迹头信息记录下来
    if (local_view_.trajectory().is_replan()) {
      latest_replan_trajectory_header_ = local_view_.trajectory().header();
    }
    // 如果最新的重新规划轨迹头具有序列号,则将其记录在控制命令的调试信息中
    if (latest_replan_trajectory_header_.has_sequence_num()) {
      debug->mutable_latest_replan_trajectory_header()->CopyFrom(
          latest_replan_trajectory_header_);
    }
  }
  // 如果当前规划的轨迹点不为空,则调用控制任务代理计算控制命令
  if (!local_view_.trajectory().trajectory_point().empty()) {
    // controller agent
    Status status_compute = control_task_agent_.ComputeControlCommand(
        &local_view_.localization(), &local_view_.chassis(),
        &local_view_.trajectory(), control_command);
    ADEBUG << "status_compute is " << status_compute;
    // 如果计算控制命令失败,记录错误信息并设置 estop_ 为 true
    if (!status_compute.ok()) {
      AERROR << "Control main function failed"
             << " with localization: "
             << local_view_.localization().ShortDebugString()
             << " with chassis: " << local_view_.chassis().ShortDebugString()
             << " with trajectory: "
             << local_view_.trajectory().ShortDebugString()
             << " with cmd: " << control_command->ShortDebugString()
             << " status:" << status_compute.error_message();
      estop_ = true;
      estop_reason_ = status_compute.error_message();
      status = status_compute;
    }
  }

  // if planning set estop, then no control process triggered
  // 如果规划停止,控制就触发不了
  if (estop_) {
    AWARN_EVERY(100) << "Estop triggered! No control core method executed!";
    // set Estop command
    control_command->set_speed(0);
    control_command->set_throttle(0);
    control_command->set_brake(FLAGS_soft_estop_brake);
    control_command->set_gear_location(Chassis::GEAR_DRIVE);
    previous_steering_command_ =
        injector_->previous_control_command_mutable()->steering_target();
    control_command->set_steering_target(previous_steering_command_);
  }
  // check signal
  if (local_view_.trajectory().decision().has_vehicle_signal()) {
    control_command->mutable_signal()->CopyFrom(
        local_view_.trajectory().decision().vehicle_signal());
  }
  return status;
}
// 核心函数Proc
bool ControlComponent::Proc() {
  const auto start_time = Clock::Now();

  chassis_reader_->Observe();
  const auto &chassis_msg = chassis_reader_->GetLatestObserved();
  // 接收不到底盘信息
  if (chassis_msg == nullptr) {
    AERROR << "Chassis msg is not ready!";
    injector_->set_control_process(false);
    return false;
  }
  OnChassis(chassis_msg);

  trajectory_reader_->Observe();
  const auto &trajectory_msg = trajectory_reader_->GetLatestObserved();
  // 接收不到轨迹信息
  if (trajectory_msg == nullptr) {
    AERROR << "planning msg is not ready!";
  } else {
    // Check if new planning data received.
    if (latest_trajectory_.header().sequence_num() !=
        trajectory_msg->header().sequence_num()) {
      OnPlanning(trajectory_msg);
    }
  }

  planning_command_status_reader_->Observe();
  const auto &planning_status_msg =
      planning_command_status_reader_->GetLatestObserved();
  if (planning_status_msg != nullptr) {
    OnPlanningCommandStatus(planning_status_msg);
    ADEBUG << "Planning command status msg is \n"
           << planning_command_status_.ShortDebugString();
  }
  injector_->set_planning_command_status(planning_command_status_);

  localization_reader_->Observe();
  const auto &localization_msg = localization_reader_->GetLatestObserved();
  // 接收不到定位消息
  if (localization_msg == nullptr) {
    AERROR << "localization msg is not ready!";
    injector_->set_control_process(false);
    return false;
  }
  OnLocalization(localization_msg);

  pad_msg_reader_->Observe();
  const auto &pad_msg = pad_msg_reader_->GetLatestObserved();
  if (pad_msg != nullptr) {
    OnPad(pad_msg);
  }

  {
    // TODO(SHU): to avoid redundent copy
    std::lock_guard<std::mutex> lock(mutex_);
    local_view_.mutable_chassis()->CopyFrom(latest_chassis_);
    local_view_.mutable_trajectory()->CopyFrom(latest_trajectory_);
    local_view_.mutable_localization()->CopyFrom(latest_localization_);
    if (pad_msg != nullptr) {
      local_view_.mutable_pad_msg()->CopyFrom(pad_msg_);
    }
  }

  // use control submodules
  if (FLAGS_use_control_submodules) {
    local_view_.mutable_header()->set_lidar_timestamp(
        local_view_.trajectory().header().lidar_timestamp());
    local_view_.mutable_header()->set_camera_timestamp(
        local_view_.trajectory().header().camera_timestamp());
    local_view_.mutable_header()->set_radar_timestamp(
        local_view_.trajectory().header().radar_timestamp());
    common::util::FillHeader(FLAGS_control_local_view_topic, &local_view_);

    const auto end_time = Clock::Now();

    // measure latency
    static apollo::common::LatencyRecorder latency_recorder(
        FLAGS_control_local_view_topic);
    latency_recorder.AppendLatencyRecord(
        local_view_.trajectory().header().lidar_timestamp(), start_time,
        end_time);

    local_view_writer_->Write(local_view_);
    return true;
  }

  if (pad_msg != nullptr) {
    ADEBUG << "pad_msg: " << pad_msg_.ShortDebugString();
    if (pad_msg_.action() == DrivingAction::RESET) {
      AINFO << "Control received RESET action!";
      estop_ = false;
      estop_reason_.clear();
    }
    pad_received_ = true;
  }

  if (FLAGS_is_control_test_mode && FLAGS_control_test_duration > 0 &&
      (start_time - init_time_).ToSecond() > FLAGS_control_test_duration) {
    AERROR << "Control finished testing. exit";
    injector_->set_control_process(false);
    return false;
  }

  injector_->set_control_process(true);

  ControlCommand control_command;

  Status status;
  // 自动驾驶模式
  if (local_view_.chassis().driving_mode() ==
      apollo::canbus::Chassis::COMPLETE_AUTO_DRIVE) {
    status = ProduceControlCommand(&control_command);
    ADEBUG << "Produce control command normal.";
  } else {
    ADEBUG << "Into reset control command.";
    ResetAndProduceZeroControlCommand(&control_command);
  }

  AERROR_IF(!status.ok()) << "Failed to produce control command:"
                          << status.error_message();

  if (pad_received_) {
    control_command.mutable_pad_msg()->CopyFrom(pad_msg_);
    pad_received_ = false;
  }

  // forward estop reason among following control frames.
  if (estop_) {
    control_command.mutable_header()->mutable_status()->set_msg(estop_reason_);
  }

  // set header
  control_command.mutable_header()->set_lidar_timestamp(
      local_view_.trajectory().header().lidar_timestamp());
  control_command.mutable_header()->set_camera_timestamp(
      local_view_.trajectory().header().camera_timestamp());
  control_command.mutable_header()->set_radar_timestamp(
      local_view_.trajectory().header().radar_timestamp());

  common::util::FillHeader(node_->Name(), &control_command);

  ADEBUG << control_command.ShortDebugString();
  if (FLAGS_is_control_test_mode) {
    ADEBUG << "Skip publish control command in test mode";
    return true;
  }

  if (fabs(control_command.debug().simple_lon_debug().vehicle_pitch()) <
      kDoubleEpsilon) {
    injector_->vehicle_state()->Update(local_view_.localization(),
                                       local_view_.chassis());
    GetVehiclePitchAngle(&control_command);
  }

  const auto end_time = Clock::Now();
  const double time_diff_ms = (end_time - start_time).ToSecond() * 1e3;
  ADEBUG << "total control time spend: " << time_diff_ms << " ms.";

  control_command.mutable_latency_stats()->set_total_time_ms(time_diff_ms);
  control_command.mutable_latency_stats()->set_total_time_exceeded(
      time_diff_ms > FLAGS_control_period * 1e3);
  ADEBUG << "control cycle time is: " << time_diff_ms << " ms.";
  status.Save(control_command.mutable_header()->mutable_status());

  // measure latency
  if (local_view_.trajectory().header().has_lidar_timestamp()) {
    static apollo::common::LatencyRecorder latency_recorder(
        FLAGS_control_command_topic);
    latency_recorder.AppendLatencyRecord(
        local_view_.trajectory().header().lidar_timestamp(), start_time,
        end_time);
  }

  // save current control command 保存当前控制命令
  injector_->Set_pervious_control_command(&control_command);
  injector_->previous_control_command_mutable()->CopyFrom(control_command);
  injector_->previous_control_debug_mutable()->CopyFrom(
      injector_->control_debug_info());
  // 发布控制命令
  control_cmd_writer_->Write(control_command);
  return true;
}
// 检查输入
Status ControlComponent::CheckInput(LocalView *local_view) {
  ADEBUG << "Received localization:"
         << local_view->localization().ShortDebugString();
  ADEBUG << "Received chassis:" << local_view->chassis().ShortDebugString();

  if (!local_view->trajectory().estop().is_estop() &&
      local_view->trajectory().trajectory_point().empty()) {
    AWARN_EVERY(100) << "planning has no trajectory point. ";
    const std::string msg =
        absl::StrCat("planning has no trajectory point. planning_seq_num:",
                     local_view->trajectory().header().sequence_num());
    return Status(ErrorCode::CONTROL_COMPUTE_ERROR, msg);
  }

  for (auto &trajectory_point :
       *local_view->mutable_trajectory()->mutable_trajectory_point()) {
    if (std::abs(trajectory_point.v()) < FLAGS_minimum_speed_resolution &&
        std::abs(trajectory_point.a()) < FLAGS_max_acceleration_when_stopped) {
      trajectory_point.set_v(0.0);
      trajectory_point.set_a(0.0);
    }
  }

  injector_->vehicle_state()->Update(local_view->localization(),
                                     local_view->chassis());

  return Status::OK();
}
// 检查时间戳
Status ControlComponent::CheckTimestamp(const LocalView &local_view) {
  if (!FLAGS_enable_input_timestamp_check || FLAGS_is_control_test_mode) {
    ADEBUG << "Skip input timestamp check by gflags.";
    return Status::OK();
  }
  double current_timestamp = Clock::NowInSeconds();
  double localization_diff =
      current_timestamp - local_view.localization().header().timestamp_sec();
  if (localization_diff >
      (FLAGS_max_localization_miss_num * FLAGS_localization_period)) {
    AERROR << "Localization msg lost for " << std::setprecision(6)
           << localization_diff << "s";
    monitor_logger_buffer_.ERROR("Localization msg lost");
    return Status(ErrorCode::CONTROL_COMPUTE_ERROR, "Localization msg timeout");
  }

  double chassis_diff =
      current_timestamp - local_view.chassis().header().timestamp_sec();
  if (chassis_diff > (FLAGS_max_chassis_miss_num * FLAGS_chassis_period)) {
    AERROR << "Chassis msg lost for " << std::setprecision(6) << chassis_diff
           << "s";
    monitor_logger_buffer_.ERROR("Chassis msg lost");
    return Status(ErrorCode::CONTROL_COMPUTE_ERROR, "Chassis msg timeout");
  }

  double trajectory_diff =
      current_timestamp - local_view.trajectory().header().timestamp_sec();
  if (trajectory_diff >
      (FLAGS_max_planning_miss_num * FLAGS_trajectory_period)) {
    AERROR << "Trajectory msg lost for " << std::setprecision(6)
           << trajectory_diff << "s";
    monitor_logger_buffer_.ERROR("Trajectory msg lost");
    return Status(ErrorCode::CONTROL_COMPUTE_ERROR, "Trajectory msg timeout");
  }
  return Status::OK();
}
// 重置控制命令
void ControlComponent::ResetAndProduceZeroControlCommand(
    ControlCommand *control_command) {
  control_command->set_throttle(0.0);
  control_command->set_steering_target(0.0);
  control_command->set_steering_rate(0.0);
  control_command->set_speed(0.0);
  control_command->set_brake(0.0);
  control_command->set_gear_location(Chassis::GEAR_DRIVE);
  control_task_agent_.Reset();
  latest_trajectory_.mutable_trajectory_point()->Clear();
  latest_trajectory_.mutable_path_point()->Clear();
  trajectory_reader_->ClearData();
}
// 获得汽车的俯仰角
void ControlComponent::GetVehiclePitchAngle(ControlCommand *control_command) {
  double vehicle_pitch = injector_->vehicle_state()->pitch() * 180 / M_PI;
  control_command->mutable_debug()
      ->mutable_simple_lon_debug()
      ->set_vehicle_pitch(vehicle_pitch + FLAGS_pitch_offset_deg);
}

}  // namespace control
}  // namespace apollo

控制全局变量配置文件见control_gflags.cc

配置加载的控制器,Apollo中modules/control/control_component/conf/pipeline.pb.txt

controller {
  name: "LAT_CONTROLLER"
  type: "LatController"
}
controller {
  name: "LON_CONTROLLER"
  type: "LonController"
}

name是用户自定义,表达清楚是什么控制器就行,type是控制器的子类名称,如果和子类名称不一致,会导致加载控制器失败。上面是先加载横向控制器,再加载纵向控制器

3 Control组件包逻辑梳理

主要是梳理Init函数和Proc函数
Init函数
主要实现ControlTaskAgent的初始化,以及control上游的相关消息的订阅

摘取Init函数里面的主要部分

bool ControlComponent::Init() {
  // 初始化控制器agent
  if (!FLAGS_is_control_ut_test_mode) {
    if (!FLAGS_use_control_submodules &&
        !control_task_agent_.Init(injector_, control_pipeline_).ok()) {
      // set controller
      ADEBUG << "original control";
      monitor_logger_buffer_.ERROR(
          "Control init controller failed! Stopping...");
      return false;
    }
  }
  // 订阅底盘信息
  chassis_reader_ =
      node_->CreateReader<Chassis>(chassis_reader_config, nullptr);
  // 订阅轨迹信息
  trajectory_reader_ =
      node_->CreateReader<ADCTrajectory>(planning_reader_config, nullptr);
  // 订阅规划命令状态
  planning_command_status_reader_ =
      node_->CreateReader<external_command::CommandStatus>(
          planning_command_status_reader_config, nullptr);
  // 订阅定位信息
  localization_reader_ = node_->CreateReader<LocalizationEstimate>(
      localization_reader_config, nullptr);
  // 订阅pad消息
  pad_msg_reader_ =
      node_->CreateReader<PadMessage>(pad_msg_reader_config, nullptr);
}

Proc函数
(1) 获取订阅消息的当前最新数据
(2) 检查订阅消息输入数据
(3) 检查订阅消息输入数据时间戳是否在容差范围内
(4) 更新车身姿态信息
(5) 进行control控制计算(调用ControlTaskAgent的ComputeControlCommand方法)
(6) 输出底盘控制指令

摘取Proc函数里面的主要部分

bool ControlComponent::Proc() {
    // 获取订阅消息的当前最新数据
	const auto &chassis_msg = chassis_reader_->GetLatestObserved();
	const auto &trajectory_msg = trajectory_reader_->GetLatestObserved();
	const auto &planning_status_msg =
      planning_command_status_reader_->GetLatestObserved();
    const auto &localization_msg = localization_reader_->GetLatestObserved();
    const auto &pad_msg = pad_msg_reader_->GetLatestObserved();
    // 检查订阅消息输入数据
    // 检查订阅消息输入数据时间戳是否在容差范围内
    // 自动驾驶模式
    if (local_view_.chassis().driving_mode() ==
        apollo::canbus::Chassis::COMPLETE_AUTO_DRIVE) {
    // 计算控制命令
      status = ProduceControlCommand(&control_command);
      ADEBUG << "Produce control command normal.";
    } else {
      ADEBUG << "Into reset control command.";
      ResetAndProduceZeroControlCommand(&control_command);
    }
    // 更新车身姿态信息
    injector_->vehicle_state()->Update(local_view_.localization(),
                                       local_view_.chassis());
    // 发布控制命令
    control_cmd_writer_->Write(control_command);
}

3 controller_task_base

之前的控制组件包讲解告一段落,如有疑惑可在评论区留言讨论

controller_task_base主要包含ControlTaskAgentControlTask定义,ControlTaskAgent用来管理ControlTask插件的加载和执行顺序,ControlTaskcontroller控制器插件的父类,Control/controller控制器插件都继承于ControlTask

control_task_agent.h

#pragma once

#include <memory>
#include <vector>

#include "modules/common_msgs/control_msgs/control_cmd.pb.h"
#include "modules/common_msgs/planning_msgs/planning.pb.h"
#include "modules/control/control_component/proto/pipeline.pb.h"

#include "cyber/plugin_manager/plugin_manager.h"
#include "modules/common/util/factory.h"
#include "modules/control/control_component/controller_task_base/common/dependency_injector.h"
#include "modules/control/control_component/controller_task_base/control_task.h"

/**
 * @namespace apollo::control
 * @brief apollo::control
 */
namespace apollo {
namespace control {

/**
 * @class ControlTaskAgent
 *
 * @brief manage all controllers declared in control config file.
 */
class ControlTaskAgent {
 public:
  /**
   * @brief 初始化 ControlTaskAgent
   * @param control_conf control configurations
   * @return Status initialization status
   */
  common::Status Init(std::shared_ptr<DependencyInjector> injector,
                      const ControlPipeline &control_pipeline);

  /**
   * @brief compute control command based on current vehicle status
   *        and target trajectory
   * @param localization vehicle location
   * @param chassis vehicle status e.g., speed, acceleration
   * @param trajectory trajectory generated by planning
   * @param cmd control command
   * @return Status computation status
   */
  // 基于当前车辆状态和目标轨迹计算控制命令
  common::Status ComputeControlCommand(
      const localization::LocalizationEstimate *localization,
      const canbus::Chassis *chassis, const planning::ADCTrajectory *trajectory,
      control::ControlCommand *cmd);

  /**
   * @brief reset ControlTaskAgent
   * @return Status reset status
   */
  // 重置ControlTaskAgent
  common::Status Reset();

 private:
  std::vector<std::shared_ptr<ControlTask>> controller_list_;
  std::shared_ptr<DependencyInjector> injector_ = nullptr;
};

}  // namespace control
}  // namespace apollo

control_task_agent.cc

#include "modules/control/control_component/controller_task_base/control_task_agent.h"

#include <utility>

#include "cyber/common/log.h"
#include "cyber/time/clock.h"
#include "modules/control/control_component/common/control_gflags.h"

namespace apollo {
namespace control {

using apollo::common::ErrorCode;
using apollo::common::Status;
using apollo::cyber::Clock;
using apollo::cyber::plugin_manager::PluginManager;
// 初始化控制器
Status ControlTaskAgent::Init(std::shared_ptr<DependencyInjector> injector,
                              const ControlPipeline &control_pipeline) {
  if (control_pipeline.controller_size() == 0) {
    AERROR << "control_pipeline is empty";
    return Status(ErrorCode::CONTROL_INIT_ERROR, "Empty control_pipeline");
  }

  injector_ = injector;
  for (int i = 0; i < control_pipeline.controller_size(); i++) {
    auto controller = PluginManager::Instance()->CreateInstance<ControlTask>(
        "apollo::control::" + control_pipeline.controller(i).type());
    if (!controller->Init(injector_).ok()) {
      AERROR << "Can not init controller " << controller->Name();
      return Status(
          ErrorCode::CONTROL_INIT_ERROR,
          "Failed to init Controller:" + control_pipeline.controller(i).name());
    }
    controller_list_.push_back(controller);
    AINFO << "Controller <" << controller->Name() << "> init done!";
  }
  return Status::OK();
}
// 计算控制命令
Status ControlTaskAgent::ComputeControlCommand(
    const localization::LocalizationEstimate *localization,
    const canbus::Chassis *chassis, const planning::ADCTrajectory *trajectory,
    control::ControlCommand *cmd) {
  for (auto &controller : controller_list_) {
    ADEBUG << "controller:" << controller->Name() << " processing ...";
    double start_timestamp = Clock::NowInSeconds();
    // 计算控制命令 (核心)
    controller->ComputeControlCommand(localization, chassis, trajectory, cmd);
    double end_timestamp = Clock::NowInSeconds();
    const double time_diff_ms = (end_timestamp - start_timestamp) * 1000;

    ADEBUG << "controller: " << controller->Name()
           << " calculation time is: " << time_diff_ms << " ms.";
    cmd->mutable_latency_stats()->add_controller_time_ms(time_diff_ms);
  }
  return Status::OK();
}

Status ControlTaskAgent::Reset() {
  for (auto &controller : controller_list_) {
    ADEBUG << "controller:" << controller->Name() << " reset...";
    controller->Reset();
  }
  return Status::OK();
}

}  // namespace control
}  // namespace apollo

ControlTask是controller控制器插件的父类,Control/controller控制器插件都继承于ControlTask

control_task.h

/**
 * @file
 * @brief Defines the Controller base class.
 */

#pragma once

#include <memory>
#include <string>

#include <cxxabi.h>

#include "modules/common_msgs/control_msgs/control_cmd.pb.h"
#include "modules/common_msgs/localization_msgs/localization.pb.h"
#include "modules/common_msgs/planning_msgs/planning.pb.h"
#include "modules/control/control_component/proto/calibration_table.pb.h"

#include "cyber/common/file.h"
#include "cyber/plugin_manager/plugin_manager.h"
#include "modules/common/status/status.h"
#include "modules/control/control_component/common/control_gflags.h"
#include "modules/control/control_component/controller_task_base/common/dependency_injector.h"

namespace apollo {
namespace control {

class ControlTask {
 public:
  ControlTask() = default;
  virtual ~ControlTask() = default;

  /**
   * @brief initialize Controller
   * @param control_conf control configurations
   * @return Status initialization status
   */
  virtual common::Status Init(std::shared_ptr<DependencyInjector> injector) = 0;

  /**
   * @brief compute control command based on current vehicle status
   *        and target trajectory
   * @param localization vehicle location
   * @param chassis vehicle status e.g., speed, acceleration
   * @param trajectory trajectory generated by planning
   * @param cmd control command
   * @return Status computation status
   */
  virtual common::Status ComputeControlCommand(
      const localization::LocalizationEstimate *localization,
      const canbus::Chassis *chassis, const planning::ADCTrajectory *trajectory,
      control::ControlCommand *cmd) = 0;

  /**
   * @brief reset Controller
   * @return Status reset status
   */
  virtual common::Status Reset() = 0;

  /**
   * @brief controller name
   * @return string controller name in string
   */
  virtual std::string Name() const = 0;

  /**
   * @brief stop controller
   */
  virtual void Stop() = 0;

 protected:
  template <typename T>
  bool LoadConfig(T *config);
  // 加载油门制动标定表
  bool LoadCalibrationTable(calibration_table *calibration_table_conf) {
    std::string calibration_table_path = FLAGS_calibration_table_file;

    if (!apollo::cyber::common::GetProtoFromFile(calibration_table_path,
                                                 calibration_table_conf)) {
      AERROR << "Load calibration table failed!";
      return false;
    }
    AINFO << "Load the calibraiton table file successfully, file path: "
          << calibration_table_path;
    return true;
  }
};

template <typename T>
bool ControlTask::LoadConfig(T *config) {
  int status;
  std::string class_name =
      abi::__cxa_demangle(typeid(*this).name(), 0, 0, &status);
  // Generate the default task config path from PluginManager.
  std::string config_path_ =
      apollo::cyber::plugin_manager::PluginManager::Instance()
          ->GetPluginConfPath<ControlTask>(class_name,
                                           "conf/controller_conf.pb.txt");

  if (!apollo::cyber::common::GetProtoFromFile(config_path_, config)) {
    AERROR << "Load config of " << class_name << " failed!";
    return false;
  }
  AINFO << "Load the [" << class_name
        << "] config file successfully, file path: " << config_path_;
  return true;
}

}  // namespace control
}  // namespace apollo

4 controller

接下来看一下具体控制器的实现
在这里插入图片描述


具体控制器讲解见下一章节

本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若转载,请注明出处:http://www.coloradmin.cn/o/1807292.html

如若内容造成侵权/违法违规/事实不符,请联系多彩编程网进行投诉反馈,一经查实,立即删除!

相关文章

计算机网络9——无线网络和移动网络2无线个人区域网 WPAN

文章目录 一、蓝牙系统二、低速 WPAN三、高速 WPAN 无线个人区域网WPAN(Wireless Personal Area Network)就是在个人工作的地方把属于个人使用的电子设备(如便携式电脑、平板电脑、便携式打印机以及蜂窝电话等)用无线技术连接起来自组网络&#xff0c;不需要使用接入点AP&#…

DP:回文串模型

一、回文子串 . - 力扣&#xff08;LeetCode&#xff09; 该题有3种解法 &#xff08;1&#xff09;中心扩展算法&#xff08;在字符串章节有介绍&#xff09;时间复杂度O&#xff08;N^2&#xff09;,空间复杂度O&#xff08;1&#xff09; &#xff08;2&#xff09;马丁车…

小冬瓜AIGC 手撕LLM 拼课

小冬瓜aigc手撕LLM学习 官方认证 手撕LLMRLHF速成班-(附赠LLM加速分布式训练超长文档&#xff09; 帮助多名同学上岸LLM方向&#xff0c;包括高校副教授&#xff0c;北美PhD&#xff0c;大厂等 课程名称【手撕LLMRLHF】 授课形式&#xff1a;在线会议直播讲解课后录播 时间&…

Nvidia的成功与竞争:CEO黄仁勋的自信与挑战

每周跟踪AI热点新闻动向和震撼发展 想要探索生成式人工智能的前沿进展吗&#xff1f;订阅我们的简报&#xff0c;深入解析最新的技术突破、实际应用案例和未来的趋势。与全球数同行一同&#xff0c;从行业内部的深度分析和实用指南中受益。不要错过这个机会&#xff0c;成为AI领…

【将xml文件转yolov5训练数据txt标签文件】连classes.txt都可以生成

将xml文件转yolov5训练数据txt标签文件 前言一、代码解析 二、使用方法总结 前言 找遍全网&#xff0c;我觉得写得最详细的就是这个博文⇨将xml文件转yolov5训练数据txt标签文件 虽然我还是没有跑成功。那个正则表达式我不会改QWQ&#xff0c;但是不妨碍我会训练ai。 最终成功…

LangChain基础知识入门

LangChain的介绍和入门 1 什么是LangChain LangChain由 Harrison Chase 创建于2022年10月&#xff0c;它是围绕LLMs&#xff08;大语言模型&#xff09;建立的一个框架&#xff0c;LLMs使用机器学习算法和海量数据来分析和理解自然语言&#xff0c;GPT3.5、GPT4是LLMs最先进的代…

架构设计-用户信息及用户相关的密码信息设计

将用户的基本信息和用户密码存放在不同的数据库表中是一种常见的安全做法&#xff0c;这种做法旨在增强数据的安全性和管理的灵活性。以下是这种做法的几个关键原因&#xff1a; 安全性增强&#xff1a; 当用户密码被单独存放在一个表中时&#xff0c;可以使用更强大的加密和哈…

kafka集成SpringBoot api编写教程

1.新建项目 用的idea是20222.1.3版本&#xff0c;没有Spring Initializr 插件&#xff0c;不能直接创建springboot项目 可以在以下网址创建项目&#xff0c;下载后解压&#xff0c;然后用idea打开项目即可 1.1 在 https://start.spring.io/ 上创建项目 1.2上传到linux&#x…

C语言 | Leetcode C语言题解之第140题单词拆分II

题目&#xff1a; 题解&#xff1a; struct Trie {int ch[26];bool flag; } trie[10001];int size;void insert(char* s, int sSize) {int add 0;for (int i 0; i < sSize; i) {int x s[i] - a;if (trie[add].ch[x] 0) {trie[add].ch[x] size;memset(trie[size].ch, 0…

读AI未来进行式笔记07量子计算

1. AI审讯技术 1.1. 发明者最初的目的是发明一种能够替代精神药物&#xff0c;为人类带来终极快乐的技术 1.1.1. 遗憾的是&#xff0c;他找到的只是通往反方向的大门 1.2. 通过非侵入式的神经电磁干扰大脑边缘系统&#xff0c;诱发受审者最…

配置 JDK 和 Android SDK

目录 一、配置JDK 1. 安装 JDK 2. JDK 环境配置 3. JDK的配置验证 二、配置 adb 和Android SDK环境 1、下载 2、配置 Android SDK 环境 一、配置JDK 1. 安装 JDK 安装链接&#xff1a;Java Downloads | Oracle 我安装的是 .zip &#xff0c;直接在指定的文件夹下解压就…

[沉迷理论]进制链表树

往期文章推荐&#xff1a; 题解之最大子矩阵-CSDN博客 洛谷P1115最大子段和[神奇的题目]-CSDN博客 &#xff08;一条神奇的分割线&#xff09; 前言 好久没有更新的我总算在百忙之中抽出时间写了篇博客。 最近总算结束了动态规划的学习&#xff0c;真的是头昏脑涨啊。 最…

论文阅读——MIRNet

项目地址&#xff1a; GitHub - swz30/MIRNet: [ECCV 2020] Learning Enriched Features for Real Image Restoration and Enhancement. SOTA results for image denoising, super-resolution, and image enhancement.GitHub - soumik12345/MIRNet: Tensorflow implementation…

【云岚到家】-day02-1-区域服务后续开发及完善

【云岚到家】-day02-1-区域服务后续开发及完善 1 区域服务后续开发1.1 添加区域服务1.1.1 接口定义1.1.1.1 接口设计1.1.1.2 接口定义-json 1.1.2 接口开发1.1.2.1 mapper1.1.2.2 service1.1.2.3 controller 1.1.3 测试 1.2 修改价格1.2.1 接口定义1.2.1.1 接口设计1.2.1.2 接口…

【RAG入门教程01】Langchian框架 v0.2介绍

LangChain 是一个开源框架&#xff0c;旨在简化使用大型语言模型 (LLM) 创建应用程序的过程。可以将其想象成一套使用高级语言工具进行搭建的乐高积木。 它对于想要构建复杂的基于语言的应用程序而又不必管理直接与语言模型交互的复杂性的开发人员特别有用。它简化了将这些模型…

动态规划(多重背包+完全背包)

P2851 [USACO06DEC] 最少的硬币 G 题解&#xff1a;从题目上看到那个有n种不同的货币&#xff0c;对于买家来说每个货币有C[ i ]个&#xff0c;是有限个数的&#xff0c;但是对于卖家来说 每个货币都是无限的&#xff0c;题目中要我们求的是买到这个物品的最小交易的货币数&…

电子设计入门教程硬件篇之集成电路IC(二)

前言&#xff1a;本文为手把手教学的电子设计入门教程硬件类的博客&#xff0c;该博客侧重针对电子设计中的硬件电路进行介绍。本篇博客将根据电子设计实战中的情况去详细讲解集成电路IC&#xff0c;这些集成电路IC包括&#xff1a;逻辑门芯片、运算放大器与电子零件。电子设计…

燃料电池汽车践行者

前言 见《氢燃料电池技术综述》 见《燃料电池工作原理详解》 见《燃料电池发电系统详解》 见《燃料电池电动汽车详解》 见《氢燃料电池汽车行业发展》 现代汽车&#xff08;中国&#xff09; 现代汽车集团&#xff0c;自1998年成立氢燃料电池研发小组以来深耕氢燃料电池技术&am…

计算机操作系统基础知识:操作系统体系结构图,操作系统的内核,大内核与微内核的区别和优缺点,时钟管理,原语

1.操作系统体系结构图&#xff1a; 2.操作系统的内核&#xff1a; 时钟管理&#xff1a;利用时钟中断实现计时功能。 原语&#xff1a;原语是一种特殊的程序&#xff0c;具有原子性。也就是说&#xff0c;这段程序运行必须一气呵成&#xff0c;不能被中断。 ubuntu、centos的…

VUE3 学习笔记(13):VUE3 下的Element-Plus基本使用

UI是页面的门面&#xff0c;一个好的UI自然令人赏心悦目&#xff1b;国人团队开发的ElementUI在众多UI中较为常见&#xff0c;因此通过介绍它的使用让大家更好的了解第三方UI的使用。 安装 Npm install element-plus --save 或 Cnpm install element-plus --save 配置 全局配置…