【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER

这篇具有很好参考价值的文章主要介绍了【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER。希望对大家有所帮助。如果存在错误或未考虑完全的地方,请大家不吝赐教,您也可以点击"举报违法"按钮提交疑问。

TASK系列解析文章

1.【Apollo学习笔记】——规划模块TASK之LANE_CHANGE_DECIDER
2.【Apollo学习笔记】——规划模块TASK之PATH_REUSE_DECIDER
3.【Apollo学习笔记】——规划模块TASK之PATH_BORROW_DECIDER
4.【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER
5.【Apollo学习笔记】——规划模块TASK之PIECEWISE_JERK_PATH_OPTIMIZER
6.【Apollo学习笔记】——规划模块TASK之PATH_ASSESSMENT_DECIDER
7.【Apollo学习笔记】——规划模块TASK之PATH_DECIDER
8.【Apollo学习笔记】——规划模块TASK之RULE_BASED_STOP_DECIDER
9.【Apollo学习笔记】——规划模块TASK之SPEED_BOUNDS_PRIORI_DECIDER&&SPEED_BOUNDS_FINAL_DECIDER
10.【Apollo学习笔记】——规划模块TASK之SPEED_HEURISTIC_OPTIMIZER
11.【Apollo学习笔记】——规划模块TASK之SPEED_DECIDER
12.【Apollo学习笔记】——规划模块TASK之PIECEWISE_JERK_SPEED_OPTIMIZER
13.【Apollo学习笔记】——规划模块TASK之PIECEWISE_JERK_NONLINEAR_SPEED_OPTIMIZER(一)
14.【Apollo学习笔记】——规划模块TASK之PIECEWISE_JERK_NONLINEAR_SPEED_OPTIMIZER(二)

前言

在Apollo星火计划学习笔记——Apollo路径规划算法原理与实践与【Apollo学习笔记】——Planning模块讲到……Stage::Process的PlanOnReferenceLine函数会依次调用task_list中的TASK,本文将会继续以LaneFollow为例依次介绍其中的TASK部分究竟做了哪些工作。由于个人能力所限,文章可能有纰漏的地方,还请批评斧正。

modules/planning/conf/scenario/lane_follow_config.pb.txt配置文件中,我们可以看到LaneFollow所需要执行的所有task。

stage_config: {
  stage_type: LANE_FOLLOW_DEFAULT_STAGE
  enabled: true
  task_type: LANE_CHANGE_DECIDER
  task_type: PATH_REUSE_DECIDER
  task_type: PATH_LANE_BORROW_DECIDER
  task_type: PATH_BOUNDS_DECIDER
  task_type: PIECEWISE_JERK_PATH_OPTIMIZER
  task_type: PATH_ASSESSMENT_DECIDER
  task_type: PATH_DECIDER
  task_type: RULE_BASED_STOP_DECIDER
  task_type: SPEED_BOUNDS_PRIORI_DECIDER
  task_type: SPEED_HEURISTIC_OPTIMIZER
  task_type: SPEED_DECIDER
  task_type: SPEED_BOUNDS_FINAL_DECIDER
  task_type: PIECEWISE_JERK_SPEED_OPTIMIZER
  # task_type: PIECEWISE_JERK_NONLINEAR_SPEED_OPTIMIZER
  task_type: RSS_DECIDER

本文将继续介绍LaneFollow的第四个TASK——PATH_BOUNDS_DECIDER

PATH_BOUNDS_DECIDER功能简介

【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶利用前几个决策器,根据相应条件,产生相应的SL边界。这里说明以下Nudge障碍物的概念——主车旁边的障碍物未完全阻挡主车,主车可以通过绕行避过障碍物(注意图中的边界)。

PATH_BOUNDS_DECIDER相关配置

modules/planning/conf/planning_config.pb.txt中有相关配置。

default_task_config: {
  task_type: PATH_BOUNDS_DECIDER
  path_bounds_decider_config {
    is_lane_borrowing: false
    is_pull_over: false
    is_extend_lane_bounds_to_include_adc: false
    pull_over_destination_to_adc_buffer: 25.0
    pull_over_destination_to_pathend_buffer: 4.0
    pull_over_road_edge_buffer: 0.15
    pull_over_approach_lon_distance_adjust_factor: 1.5
  }
}

modules/planning/proto/task_config.proto中也有相关参数

// PathBoundsDeciderConfig

message PathBoundsDeciderConfig {
  optional bool is_lane_borrowing = 1;
  optional bool is_pull_over = 2;
  // not search pull-over position if the destination is within this distance
  // from ADC
  optional double pull_over_destination_to_adc_buffer = 3 [default = 25.0];
  // not search pull-over position if the destination is within this distance to
  // path-end
  optional double pull_over_destination_to_pathend_buffer = 4 [default = 10.0];
  // disquality a pull-over position if the available path boundary's edge is
  // not within this distance from the road edge
  optional double pull_over_road_edge_buffer = 5 [default = 0.15];
  optional double pull_over_approach_lon_distance_adjust_factor = 6
      [default = 1.5];
  optional double adc_buffer_coeff = 7 [default = 1.0];
  optional bool is_extend_lane_bounds_to_include_adc = 8 [default = true];
}

数据结构

// modules/planning/tasks/deciders/path_bounds_decider/path_bounds_decider.cc
namespace {
// PathBoundPoint contains: (s, l_min, l_max).路径边界点
using PathBoundPoint = std::tuple<double, double, double>;
// PathBound contains a vector of PathBoundPoints.路径边界
using PathBound = std::vector<PathBoundPoint>;
// ObstacleEdge contains: (is_start_s, s, l_min, l_max, obstacle_id).障碍物的边
using ObstacleEdge = std::tuple<int, double, double, double, std::string>;
}  // namespace

参数设置

// modules/planning/tasks/deciders/path_bounds_decider/path_bounds_decider.h
// s方向的距离
constexpr double kPathBoundsDeciderHorizon = 100.0;
// s方向的间隔
constexpr double kPathBoundsDeciderResolution = 0.5;
// Lane宽度
constexpr double kDefaultLaneWidth = 5.0;
// Road的道路
constexpr double kDefaultRoadWidth = 20.0;
// TODO(all): Update extra tail point base on vehicle speed.
constexpr int kNumExtraTailBoundPoint = 20;
constexpr double kPulloverLonSearchCoeff = 1.5;
constexpr double kPulloverLatSearchCoeff = 1.25;

PATH_BOUNDS_DECIDER总体流程

总体流程依然先看Process部分
【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶
Process方法中,分四种场景对路径边界进行计算,按照处理的顺序分别是:fallbackpull-overlane-changeregular。 其中regular场景根据是否借道又分为LEFT_BORROW, NO_BORROW, RIGHT_BORROW

fallback场景的path bounds一定会生成,另外三种看情况,都是需要if判断。

Status PathBoundsDecider::Process(
    Frame* const frame, ReferenceLineInfo* const reference_line_info) {
  // Sanity checks.
  CHECK_NOTNULL(frame);
  CHECK_NOTNULL(reference_line_info);
  // 如果道路重用置位,则跳过PathBoundsDecider
  // Skip the path boundary decision if reusing the path.
  if (FLAGS_enable_skip_path_tasks && reference_line_info->path_reusable()) {
    return Status::OK();
  }

  std::vector<PathBoundary> candidate_path_boundaries;
  // const TaskConfig& config = Decider::config_;

  // Initialization.
  // 用规划起始点自车道的lane width去初始化 path boundary
  // 如果无法获取规划起始点的道路宽度,则用默认值目前是5m,来初始化
  InitPathBoundsDecider(*frame, *reference_line_info);

  // Generate the fallback path boundary.
  // 生成fallback_path_bound,无论何种场景都会生成fallback_path_bound
  // 生成fallback_path_bound时,会考虑借道场景,向哪个方向借道,对应方向的path_bound会叠加这个方向相邻车道宽度
  PathBound fallback_path_bound;
  // 生成fallback_path_bound,
  // 首先获取当前车道的宽度
  // 再根据是否发生借道的决策,获取相邻车道的宽度(包括对向车道以及同向相邻车道);
  // 再根据ADC的位置和速度计算边界;
  // 最后更新边界。
  Status ret =
      GenerateFallbackPathBound(*reference_line_info, &fallback_path_bound);
  if (!ret.ok()) {
    ADEBUG << "Cannot generate a fallback path bound.";
    return Status(ErrorCode::PLANNING_ERROR, ret.error_message());
  }
  if (fallback_path_bound.empty()) {
    const std::string msg = "Failed to get a valid fallback path boundary";
    AERROR << msg;
    return Status(ErrorCode::PLANNING_ERROR, msg);
  }
  if (!fallback_path_bound.empty()) {
    CHECK_LE(adc_frenet_l_, std::get<2>(fallback_path_bound[0]));
    CHECK_GE(adc_frenet_l_, std::get<1>(fallback_path_bound[0]));
  }
  // Update the fallback path boundary into the reference_line_info.
  std::vector<std::pair<double, double>> fallback_path_bound_pair;
  for (size_t i = 0; i < fallback_path_bound.size(); ++i) {
    fallback_path_bound_pair.emplace_back(std::get<1>(fallback_path_bound[i]),
                                          std::get<2>(fallback_path_bound[i]));
  }
  //生成fallback边界,存入candidate_path_boundaries,并标记为“fallback”
  candidate_path_boundaries.emplace_back(std::get<0>(fallback_path_bound[0]),
                                         kPathBoundsDeciderResolution,
                                         fallback_path_bound_pair);
  candidate_path_boundaries.back().set_label("fallback");
  // 根据场景生成另外一条path_bound
  // 依次判断是否处于pull_over、lane_change、regular;当处于其中一个场景,计算对应的path_bound并返回结果;
  // 即以上3种场景,只会选一种生成对应的根据场景生成另外一条path_bound
 
  // 判断是否为pull_over_status
  // If pull-over is requested, generate pull-over path boundary.
  auto* pull_over_status = injector_->planning_context()
                               ->mutable_planning_status()
                               ->mutable_pull_over();
  const bool plan_pull_over_path = pull_over_status->plan_pull_over_path();
  if (plan_pull_over_path) {
    PathBound pull_over_path_bound;
    Status ret = GeneratePullOverPathBound(*frame, *reference_line_info,
                                           &pull_over_path_bound);
    if (!ret.ok()) {
      AWARN << "Cannot generate a pullover path bound, do regular planning.";
    } else {
      ACHECK(!pull_over_path_bound.empty());
      CHECK_LE(adc_frenet_l_, std::get<2>(pull_over_path_bound[0]));
      CHECK_GE(adc_frenet_l_, std::get<1>(pull_over_path_bound[0]));

      // Update the fallback path boundary into the reference_line_info.
      std::vector<std::pair<double, double>> pull_over_path_bound_pair;
      for (size_t i = 0; i < pull_over_path_bound.size(); ++i) {
        pull_over_path_bound_pair.emplace_back(
            std::get<1>(pull_over_path_bound[i]),
            std::get<2>(pull_over_path_bound[i]));
      }
      candidate_path_boundaries.emplace_back(
          std::get<0>(pull_over_path_bound[0]), kPathBoundsDeciderResolution,
          pull_over_path_bound_pair);
      candidate_path_boundaries.back().set_label("regular/pullover");

      reference_line_info->SetCandidatePathBoundaries(
          std::move(candidate_path_boundaries));
      ADEBUG << "Completed pullover and fallback path boundaries generation.";

      // set debug info in planning_data
      auto* pull_over_debug = reference_line_info->mutable_debug()
                                  ->mutable_planning_data()
                                  ->mutable_pull_over();
      pull_over_debug->mutable_position()->CopyFrom(
          pull_over_status->position());
      pull_over_debug->set_theta(pull_over_status->theta());
      pull_over_debug->set_length_front(pull_over_status->length_front());
      pull_over_debug->set_length_back(pull_over_status->length_back());
      pull_over_debug->set_width_left(pull_over_status->width_left());
      pull_over_debug->set_width_right(pull_over_status->width_right());

      return Status::OK();
    }
  }
  // 判断是否满足lane_change场景
  // 首先FLAGS_enable_smarter_lane_change 要置位
  // 通过判断当前reference_line 是否为目标车道,来进行判断
  // If it's a lane-change reference-line, generate lane-change path boundary.
  if (FLAGS_enable_smarter_lane_change &&
      reference_line_info->IsChangeLanePath()) {
    PathBound lanechange_path_bound;
    Status ret = GenerateLaneChangePathBound(*reference_line_info,
                                             &lanechange_path_bound);
    if (!ret.ok()) {
      ADEBUG << "Cannot generate a lane-change path bound.";
      return Status(ErrorCode::PLANNING_ERROR, ret.error_message());
    }
    if (lanechange_path_bound.empty()) {
      const std::string msg = "Failed to get a valid fallback path boundary";
      AERROR << msg;
      return Status(ErrorCode::PLANNING_ERROR, msg);
    }

    // disable this change when not extending lane bounds to include adc
    if (config_.path_bounds_decider_config()
            .is_extend_lane_bounds_to_include_adc()) {
      CHECK_LE(adc_frenet_l_, std::get<2>(lanechange_path_bound[0]));
      CHECK_GE(adc_frenet_l_, std::get<1>(lanechange_path_bound[0]));
    }
    // Update the fallback path boundary into the reference_line_info.
    std::vector<std::pair<double, double>> lanechange_path_bound_pair;
    for (size_t i = 0; i < lanechange_path_bound.size(); ++i) {
      lanechange_path_bound_pair.emplace_back(
          std::get<1>(lanechange_path_bound[i]),
          std::get<2>(lanechange_path_bound[i]));
    }
    candidate_path_boundaries.emplace_back(
        std::get<0>(lanechange_path_bound[0]), kPathBoundsDeciderResolution,
        lanechange_path_bound_pair);
    candidate_path_boundaries.back().set_label("regular/lanechange");
    RecordDebugInfo(lanechange_path_bound, "", reference_line_info);
    reference_line_info->SetCandidatePathBoundaries(
        std::move(candidate_path_boundaries));
    ADEBUG << "Completed lanechange and fallback path boundaries generation.";
    return Status::OK();
  }

  // Generate regular path boundaries.
  // 生成regular path boundaries
  std::vector<LaneBorrowInfo> lane_borrow_info_list;
  lane_borrow_info_list.push_back(LaneBorrowInfo::NO_BORROW);

  if (reference_line_info->is_path_lane_borrow()) {
    const auto& path_decider_status =
        injector_->planning_context()->planning_status().path_decider();
    for (const auto& lane_borrow_direction :
         path_decider_status.decided_side_pass_direction()) {
      if (lane_borrow_direction == PathDeciderStatus::LEFT_BORROW) {
        lane_borrow_info_list.push_back(LaneBorrowInfo::LEFT_BORROW);
      } else if (lane_borrow_direction == PathDeciderStatus::RIGHT_BORROW) {
        lane_borrow_info_list.push_back(LaneBorrowInfo::RIGHT_BORROW);
      }
    }
  }

  // Try every possible lane-borrow option:
  // PathBound regular_self_path_bound;
  // bool exist_self_path_bound = false;
  for (const auto& lane_borrow_info : lane_borrow_info_list) {
    PathBound regular_path_bound;
    std::string blocking_obstacle_id = "";
    std::string borrow_lane_type = "";
    Status ret = GenerateRegularPathBound(
        *reference_line_info, lane_borrow_info, &regular_path_bound,
        &blocking_obstacle_id, &borrow_lane_type);
    if (!ret.ok()) {
      continue;
    }
    if (regular_path_bound.empty()) {
      continue;
    }
    // disable this change when not extending lane bounds to include adc
    if (config_.path_bounds_decider_config()
            .is_extend_lane_bounds_to_include_adc()) {
      CHECK_LE(adc_frenet_l_, std::get<2>(regular_path_bound[0]));
      CHECK_GE(adc_frenet_l_, std::get<1>(regular_path_bound[0]));
    }

    // Update the path boundary into the reference_line_info.
    std::vector<std::pair<double, double>> regular_path_bound_pair;
    for (size_t i = 0; i < regular_path_bound.size(); ++i) {
      regular_path_bound_pair.emplace_back(std::get<1>(regular_path_bound[i]),
                                           std::get<2>(regular_path_bound[i]));
    }
    candidate_path_boundaries.emplace_back(std::get<0>(regular_path_bound[0]),
                                           kPathBoundsDeciderResolution,
                                           regular_path_bound_pair);
    std::string path_label = "";
    switch (lane_borrow_info) {
      case LaneBorrowInfo::LEFT_BORROW:
        path_label = "left";
        break;
      case LaneBorrowInfo::RIGHT_BORROW:
        path_label = "right";
        break;
      default:
        path_label = "self";
        // exist_self_path_bound = true;
        // regular_self_path_bound = regular_path_bound;
        break;
    }
    // RecordDebugInfo(regular_path_bound, "", reference_line_info);
    candidate_path_boundaries.back().set_label(
        absl::StrCat("regular/", path_label, "/", borrow_lane_type));
    candidate_path_boundaries.back().set_blocking_obstacle_id(
        blocking_obstacle_id);
  }

  // Remove redundant boundaries.
  // RemoveRedundantPathBoundaries(&candidate_path_boundaries);

  // Success
  reference_line_info->SetCandidatePathBoundaries(
      std::move(candidate_path_boundaries));
  ADEBUG << "Completed regular and fallback path boundaries generation.";
  return Status::OK();
}

InitPathBoundsDecider

InitPathBoundsDecider初始化planning_start_point、获取ADC S/L坐标信息、获取ADC当前所在的车道线宽度。

void PathBoundsDecider::InitPathBoundsDecider(
    const Frame& frame, const ReferenceLineInfo& reference_line_info) {
  const ReferenceLine& reference_line = reference_line_info.reference_line();
  common::TrajectoryPoint planning_start_point = frame.PlanningStartPoint();
  if (FLAGS_use_front_axe_center_in_path_planning) {
    planning_start_point =
        InferFrontAxeCenterFromRearAxeCenter(planning_start_point);
  }
  ADEBUG << "Plan at the starting point: x = "
         << planning_start_point.path_point().x()
         << ", y = " << planning_start_point.path_point().y()
         << ", and angle = " << planning_start_point.path_point().theta();

  // Initialize some private variables.
  // ADC s/l info.
  auto adc_sl_info = reference_line.ToFrenetFrame(planning_start_point);
  adc_frenet_s_ = adc_sl_info.first[0];
  adc_frenet_l_ = adc_sl_info.second[0];
  adc_frenet_sd_ = adc_sl_info.first[1];
  adc_frenet_ld_ = adc_sl_info.second[1] * adc_frenet_sd_;
  double offset_to_map = 0.0;
  reference_line.GetOffsetToMap(adc_frenet_s_, &offset_to_map);
  adc_l_to_lane_center_ = adc_frenet_l_ + offset_to_map;

  // ADC's lane width.
  double lane_left_width = 0.0;
  double lane_right_width = 0.0;
  if (!reference_line.GetLaneWidth(adc_frenet_s_, &lane_left_width,
                                   &lane_right_width)) {
    AWARN << "Failed to get lane width at planning start point.";
    adc_lane_width_ = kDefaultLaneWidth;
  } else {
    adc_lane_width_ = lane_left_width + lane_right_width;
  }
}

1. fallback

无论当前处于何种场景,都会调用GenerateFallbackPathBound() 来生成备用的path bound,在计算FallbackPathBound时不考虑障碍物边界,仅使用道路边界,并考虑借道信息来进行计算。
【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶
fallback部分主要包含两个函数,InitPathBoundaryGetBoundaryFromLanesAndADC。fallback是备选的方案,只考虑adc信息和静态道路信息。

GenerateFallbackPathBound

Status PathBoundsDecider::GenerateFallbackPathBound(
    const ReferenceLineInfo& reference_line_info, PathBound* const path_bound) {
  // 1. Initialize the path boundaries to be an indefinitely large area.
  if (!InitPathBoundary(reference_line_info, path_bound)) {...
  // PathBoundsDebugString(*path_bound);

  // 2. Decide a rough boundary based on lane info and ADC's position
  std::string dummy_borrow_lane_type;
  if (!GetBoundaryFromLanesAndADC(reference_line_info,
                                  LaneBorrowInfo::NO_BORROW, 0.5, path_bound,
                                  &dummy_borrow_lane_type, true)) {...
  // PathBoundsDebugString(*path_bound);
  ADEBUG << "Completed generating fallback path boundaries.";
  return Status::OK();
}
InitPathBoundary

InitPathBoundary从adc当前位置开始,以0.5m为间隔取点,直到终点,将 [左, 右] 边界设置为double的 [lowerst, max]。

bool PathBoundsDecider::InitPathBoundary(
  ...
  // Starting from ADC's current position, increment until the horizon, and
  // set lateral bounds to be infinite at every spot.
  // 从adc当前位置开始,以0.5m为间隔取点,直到终点,将 [左, 右] 边界设置为double的 [lowerst, max]
  for (double curr_s = adc_frenet_s_;
       curr_s < std::fmin(adc_frenet_s_ +
                              std::fmax(kPathBoundsDeciderHorizon,
                                        reference_line_info.GetCruiseSpeed() *
                                            FLAGS_trajectory_time_length),
                          reference_line.Length());
       curr_s += kPathBoundsDeciderResolution) {
    path_bound->emplace_back(curr_s, std::numeric_limits<double>::lowest(),
                             std::numeric_limits<double>::max());
  }
...}
GetBoundaryFromLanesAndADC

GetBoundaryFromLanesAndADC首先获取当前车道的宽度;再根据是否发生借道的决策,获取相邻车道的宽度(包括对向车道以及同向相邻车道);再根据ADC的位置和速度计算边界;最后更新边界。

// TODO(jiacheng): this function is to be retired soon.
bool PathBoundsDecider::GetBoundaryFromLanesAndADC(
  ...
  for (size_t i = 0; i < path_bound->size(); ++i) {
    double curr_s = std::get<0>((*path_bound)[i]);
    // 1. Get the current lane width at current point.获取当前点车道的宽度
    if (!reference_line.GetLaneWidth(curr_s, &curr_lane_left_width,
                                     &curr_lane_right_width)) {
      AWARN << "Failed to get lane width at s = " << curr_s;
      curr_lane_left_width = past_lane_left_width;
      curr_lane_right_width = past_lane_right_width;
    } else {...}

    // 2. Get the neighbor lane widths at the current point.获取当前点相邻车道的宽度
    double curr_neighbor_lane_width = 0.0;
    if (CheckLaneBoundaryType(reference_line_info, curr_s, lane_borrow_info)) {
      hdmap::Id neighbor_lane_id;
      if (lane_borrow_info == LaneBorrowInfo::LEFT_BORROW) {
        // 借左车道
        ...
      } else if (lane_borrow_info == LaneBorrowInfo::RIGHT_BORROW) {
        // 借右车道
        ...
      }
    }

    // 3. 根据道路宽度,adc的位置和速度计算合适的边界。
    static constexpr double kMaxLateralAccelerations = 1.5;
    double offset_to_map = 0.0;
    reference_line.GetOffsetToMap(curr_s, &offset_to_map);

    double ADC_speed_buffer = (adc_frenet_ld_ > 0 ? 1.0 : -1.0) *
                              adc_frenet_ld_ * adc_frenet_ld_ /
                              kMaxLateralAccelerations / 2.0;
    // 向左车道借到,左边界会变成左侧车道左边界
    double curr_left_bound_lane =
        curr_lane_left_width + (lane_borrow_info == LaneBorrowInfo::LEFT_BORROW
                                    ? curr_neighbor_lane_width
                                    : 0.0);
    // 和上面类似
    double curr_right_bound_lane =
        -curr_lane_right_width -
        (lane_borrow_info == LaneBorrowInfo::RIGHT_BORROW
             ? curr_neighbor_lane_width
             : 0.0);

    double curr_left_bound = 0.0;  // 左边界
    double curr_right_bound = 0.0;  // 右边界
    // 计算左边界和右边界
    if (config_.path_bounds_decider_config()
            .is_extend_lane_bounds_to_include_adc() ||
        is_fallback_lanechange) {
      // extend path bounds to include ADC in fallback or change lane path
      // bounds.
      double curr_left_bound_adc =
          std::fmax(adc_l_to_lane_center_,
                    adc_l_to_lane_center_ + ADC_speed_buffer) +
          GetBufferBetweenADCCenterAndEdge() + ADC_buffer;
      curr_left_bound =
          std::fmax(curr_left_bound_lane, curr_left_bound_adc) - offset_to_map;

      double curr_right_bound_adc =
          std::fmin(adc_l_to_lane_center_,
                    adc_l_to_lane_center_ + ADC_speed_buffer) -
          GetBufferBetweenADCCenterAndEdge() - ADC_buffer;
      curr_right_bound =
          std::fmin(curr_right_bound_lane, curr_right_bound_adc) -
          offset_to_map;
    } else {
      curr_left_bound = curr_left_bound_lane - offset_to_map;
      curr_right_bound = curr_right_bound_lane - offset_to_map;
    }

    // 4. 更新边界.
    if (!UpdatePathBoundaryWithBuffer(i, curr_left_bound, curr_right_bound,
                                      path_bound, is_left_lane_boundary,
                                      is_right_lane_boundary)) {
      path_blocked_idx = static_cast<int>(i);
    }
... }

PS:
涉及到的几个参数:
(1)ADC_speed_buffer,我的理解是车辆以最大横向加速度将横向速度降为0所经过的距离。 l a t = V l a t 2 2 a lat=\frac {{V_{lat}}^2} {2a} lat=2aVlat2

    double ADC_speed_buffer = (adc_frenet_ld_ > 0 ? 1.0 : -1.0) *
                              adc_frenet_ld_ * adc_frenet_ld_ /
                              kMaxLateralAccelerations / 2.0;

(2) curr_left_bound_lanecurr_right_bound_lane基于车道线的左右边界,如果产生变道,则加上相邻车道的宽度。

    double curr_left_bound_lane =
        curr_lane_left_width + (lane_borrow_info == LaneBorrowInfo::LEFT_BORROW
                                    ? curr_neighbor_lane_width
                                    : 0.0);

    double curr_right_bound_lane =
        -curr_lane_right_width -
        (lane_borrow_info == LaneBorrowInfo::RIGHT_BORROW
             ? curr_neighbor_lane_width
             : 0.0);

(3)curr_left_bound_adc基于自车产生的边界。根据ADC_speed_buffer,ADC_buffer(0.5,可以认为是横向安全区间),半车宽,adc_l_to_lane_center_(自车相对于车道中心的偏移量)计算。

      double curr_left_bound_adc =
          std::fmax(adc_l_to_lane_center_,
                    adc_l_to_lane_center_ + ADC_speed_buffer) +
          GetBufferBetweenADCCenterAndEdge() + ADC_buffer;

计算示意图大致如下:
【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶
(4)offset_to_map:相对地图的偏移量。

UpdatePathBoundaryWithBuffer

UpdatePathBoundaryWithBuffer首先计算缓冲边界(实际意义就是将左右边界各收缩半个车宽),若(l_min > l_max),即右边界大于左边界,则发生block,不更新。

bool PathBoundsDecider::UpdatePathBoundaryWithBuffer(
    size_t idx, double left_bound, double right_bound,
    PathBound* const path_boundaries, bool is_left_lane_bound,
    bool is_right_lane_bound) {
  // substract vehicle width when bound does not come from the lane boundary
  // 计算缓冲边界,默认是1.0
  const double default_adc_buffer_coeff = 1.0;
  double left_adc_buffer_coeff =
      (is_left_lane_bound
           ? config_.path_bounds_decider_config().adc_buffer_coeff()
           : default_adc_buffer_coeff);
  double right_adc_buffer_coeff =
      (is_right_lane_bound
           ? config_.path_bounds_decider_config().adc_buffer_coeff()
           : default_adc_buffer_coeff);

  // Update the right bound (l_min):
  double new_l_min =
      std::fmax(std::get<1>((*path_boundaries)[idx]),
                right_bound + right_adc_buffer_coeff *
                                  GetBufferBetweenADCCenterAndEdge());
  // Update the left bound (l_max):
  double new_l_max = std::fmin(
      std::get<2>((*path_boundaries)[idx]),
      left_bound - left_adc_buffer_coeff * GetBufferBetweenADCCenterAndEdge());

  // Check if ADC is blocked.
  // If blocked, don't update anything, return false.
  if (new_l_min > new_l_max) {
    ADEBUG << "Path is blocked at idx = " << idx;
    return false;
  }
  // Otherwise, update path_boundaries and center_line; then return true.
  std::get<1>((*path_boundaries)[idx]) = new_l_min;
  std::get<2>((*path_boundaries)[idx]) = new_l_max;
  return true;
}
TrimPathBounds

TrimPathBounds截断block之后的边界点。

2. pull over

【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶

GeneratePullOverPathBound

核心逻辑在GeneratePullOverPathBound中。
InitPathBoundary在fallback部分已经讨论过了,接着来看GetBoundaryFromRoads

GetBoundaryFromRoads

GetBoundaryFromRoads主要完成:获取道路边界;更新道路边界;裁剪block之后的边界点。

GetBoundaryFromRoads根据道路信息确定一个大致的边界。返回的边界是相对于车道中心(而不是reference_line)的,尽管在大多数情况下reference_line与车道中心的偏差可以忽略不计。

UpdatePullOverBoundaryByLaneBoundary

之后调用UpdatePullOverBoundaryByLaneBoundary更新pull over的边界。pull over有两种状态:PULL_OVEREMERGENCY_PULL_OVER
对于PULL_OVER,选取车道左边界作为边界。对于EMERGENCY_PULL_OVER,选取左/右车道边界作为边界,因为可能会在对向车道停车。

GetBoundaryFromStaticObstacles

根据静态障碍物对边界进行调整。它将确保边界不包含任何静态障碍,这样之后优化时生成的路径就不会与任何静态障碍碰撞。

【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶

SearchPullOverPosition

它的目的是搜索停车点,为了能够容纳车的尺寸,因为要先搜索可以停车的区域,然后在该区域内取一点作为停车点。搜索区域时,要先确定一个端点,然后向前或向后考察另一个端点,以及考察两端点之间的区域是否符合要求。

搜索pull over位置的过程:

  • 根据pull_over_status.pull_over_type()判断是前向搜索(pull
    over开头第一个点),还是后向搜索(pull over末尾后一个点)
  • 两层循环,外层控制搜索的索引idx,内层控制进一步的索引(前向idx+1,后向idx-1)。
  • 根据内外两层循环的索引,判断搜索到的空间是否满足宽度和长度要求,判断是否可以pull over
bool PathBoundsDecider::SearchPullOverPosition(
    const Frame& frame, const ReferenceLineInfo& reference_line_info,
    const std::vector<std::tuple<double, double, double>>& path_bound,
    std::tuple<double, double, double, int>* const pull_over_configuration) {
  const auto& pull_over_status =
      injector_->planning_context()->planning_status().pull_over();

  // 搜索方向,默认前向搜索
  bool search_backward = false;  // search FORWARD by default

  double pull_over_s = 0.0;
  if (pull_over_status.pull_over_type() ==
      PullOverStatus::EMERGENCY_PULL_OVER) {...}

  int idx = 0;
  if (search_backward) {
    // 后向搜索,定位pull over末尾的一个点.
    idx = static_cast<int>(path_bound.size()) - 1;
    while (idx >= 0 && std::get<0>(path_bound[idx]) > pull_over_s) {
      --idx;
    }
  } else {
    // 前向搜索,定位emergency pull over开头后的第一个点.
    while (idx < static_cast<int>(path_bound.size()) &&
           std::get<0>(path_bound[idx]) < pull_over_s) {
      ++idx;
    }
  }
  
  // 为pull over搜索到一个可行的位置,主要是确定该区域的宽度和长度
  const double pull_over_space_length =
      kPulloverLonSearchCoeff *
          VehicleConfigHelper::GetConfig().vehicle_param().length() -
      FLAGS_obstacle_lon_start_buffer - FLAGS_obstacle_lon_end_buffer;
  const double pull_over_space_width =
      (kPulloverLatSearchCoeff - 1.0) *
      VehicleConfigHelper::GetConfig().vehicle_param().width();
  const double adc_half_width =
      VehicleConfigHelper::GetConfig().vehicle_param().width() / 2.0;

  // 2. Find a window that is close to road-edge.
  /*
    这里用了内外两层循环进行搜索,外层循环控制搜索的开始的端点idx。
    内层控制另一个端点。根据找到的两个端点,判断区域是否可以pull over
  */
  bool has_a_feasible_window = false;
  while ((search_backward && idx >= 0 &&
          std::get<0>(path_bound[idx]) - std::get<0>(path_bound.front()) >
              pull_over_space_length) ||
         (!search_backward && idx < static_cast<int>(path_bound.size()) &&
          std::get<0>(path_bound.back()) - std::get<0>(path_bound[idx]) >
              pull_over_space_length)) {

    while ((search_backward && j >= 0 &&
            std::get<0>(path_bound[idx]) - std::get<0>(path_bound[j]) <
                pull_over_space_length) ||
           (!search_backward && j < static_cast<int>(path_bound.size()) &&
            std::get<0>(path_bound[j]) - std::get<0>(path_bound[idx]) <
                pull_over_space_length)) {...}
    
    // 找到可行区域,获取停车区域的位置和姿态
    if (is_feasible_window) {
    ...
    break;}
    ...}  // 外层while
...
}

3. lane change

【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶
核心逻辑在GenerateLaneChangePathBound函数中,与前面的场景计算流程大同小异。主要核心部分在GetBoundaryFromLaneChangeForbiddenZone部分。

GenerateLaneChangePathBound

Status PathBoundsDecider::GenerateLaneChangePathBound(
    const ReferenceLineInfo& reference_line_info,
    std::vector<std::tuple<double, double, double>>* const path_bound) {
  // 1.初始化,和前面的步骤类似
  if (!InitPathBoundary(reference_line_info, path_bound)) {...}


  // 2. 根据道路和adc的信息获取一个大致的路径边界
  std::string dummy_borrow_lane_type;
  if (!GetBoundaryFromLanesAndADC(reference_line_info,
                                  LaneBorrowInfo::NO_BORROW, 0.1, path_bound,
                                  &dummy_borrow_lane_type, true)) {...}
 

  // 3. Remove the S-length of target lane out of the path-bound.
  GetBoundaryFromLaneChangeForbiddenZone(reference_line_info, path_bound);


  // 根据静态障碍物调整边界.
  if (!GetBoundaryFromStaticObstacles(reference_line_info.path_decision(),
                                      path_bound, &blocking_obstacle_id)) {...}
  ...
}
GetBoundaryFromLaneChangeForbiddenZone

GetBoundaryFromLaneChangeForbiddenZone函数运行流程如下:

  • 如果当前位置可以变道,则直接变道
  • 如果有一个lane-change的起点,则直接使用它
  • 逐个检查变道前的点的边界,改变边界的值(如果已经过了变道点,则返回)
void PathBoundsDecider::GetBoundaryFromLaneChangeForbiddenZone(
    const ReferenceLineInfo& reference_line_info, PathBound* const path_bound) {
  // Sanity checks.
  CHECK_NOTNULL(path_bound);
  const ReferenceLine& reference_line = reference_line_info.reference_line();

  // If there is a pre-determined lane-change starting position, then use it;
  // otherwise, decide one.
  auto* lane_change_status = injector_->planning_context()
                                 ->mutable_planning_status()
                                 ->mutable_change_lane();
  // 1.当前位置直接变道。
  if (lane_change_status->is_clear_to_change_lane()) {
    ADEBUG << "Current position is clear to change lane. No need prep s.";
    lane_change_status->set_exist_lane_change_start_position(false);
    return;
  }
  double lane_change_start_s = 0.0;
  // 2.如果已经有一个lane-change的起点,就直接使用它,否则再找一个
  if (lane_change_status->exist_lane_change_start_position()) {
    common::SLPoint point_sl;
    reference_line.XYToSL(lane_change_status->lane_change_start_position(),
                          &point_sl);
    lane_change_start_s = point_sl.s();
  } else {
    // TODO(jiacheng): train ML model to learn this.
    // 设置为adc前方一段距离为变道起始点
    lane_change_start_s = FLAGS_lane_change_prepare_length + adc_frenet_s_;
    
    // Update the decided lane_change_start_s into planning-context.
    // 更新变道起始点的信息
    common::SLPoint lane_change_start_sl;
    lane_change_start_sl.set_s(lane_change_start_s);
    lane_change_start_sl.set_l(0.0);
    common::math::Vec2d lane_change_start_xy;
    reference_line.SLToXY(lane_change_start_sl, &lane_change_start_xy);
    lane_change_status->set_exist_lane_change_start_position(true);
    lane_change_status->mutable_lane_change_start_position()->set_x(
        lane_change_start_xy.x());
    lane_change_status->mutable_lane_change_start_position()->set_y(
        lane_change_start_xy.y());
  }

  // Remove the target lane out of the path-boundary, up to the decided S.
  if (lane_change_start_s < adc_frenet_s_) {
    // If already passed the decided S, then return.
    // lane_change_status->set_exist_lane_change_start_position(false);
    return;
  }
  // 逐个检查变道前的点的边界,改变边界的值
  for (size_t i = 0; i < path_bound->size(); ++i) {
    double curr_s = std::get<0>((*path_bound)[i]);
    if (curr_s > lane_change_start_s) {
      break;
    }
    double curr_lane_left_width = 0.0;
    double curr_lane_right_width = 0.0;
    double offset_to_map = 0.0;
    reference_line.GetOffsetToMap(curr_s, &offset_to_map);
    if (reference_line.GetLaneWidth(curr_s, &curr_lane_left_width,
                                    &curr_lane_right_width)) {
      double offset_to_lane_center = 0.0;
      reference_line.GetOffsetToMap(curr_s, &offset_to_lane_center);
      curr_lane_left_width += offset_to_lane_center;
      curr_lane_right_width -= offset_to_lane_center;
    }
    curr_lane_left_width -= offset_to_map;
    curr_lane_right_width += offset_to_map;

    std::get<1>((*path_bound)[i]) =
        adc_frenet_l_ > curr_lane_left_width   //右变道已跨过车道线,设定右界限
            ? curr_lane_left_width + GetBufferBetweenADCCenterAndEdge()
            : std::get<1>((*path_bound)[i]);
    std::get<1>((*path_bound)[i]) =
        std::fmin(std::get<1>((*path_bound)[i]), adc_frenet_l_ - 0.1);
    std::get<2>((*path_bound)[i]) =
        adc_frenet_l_ < -curr_lane_right_width
            ? -curr_lane_right_width - GetBufferBetweenADCCenterAndEdge()
            : std::get<2>((*path_bound)[i]);  //左变道已跨过车道线,设定左界限
    std::get<2>((*path_bound)[i]) =
        std::fmax(std::get<2>((*path_bound)[i]), adc_frenet_l_ + 0.1);
  }
}

PS:变道时是以相邻车道的参考线作为参考线。以向右变道为例,变道前(指跨过车道线线前adc_frenet_l_ > curr_lane_left_width),使用原先的path bound,不给右侧留空间;变道之后,再进行更新。

4. regular

【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER,Apollo,planning,算法,决策规划,apollo,自动驾驶
代码流程如下:

Status PathBoundsDecider::GenerateRegularPathBound(
    const ReferenceLineInfo& reference_line_info,
    const LaneBorrowInfo& lane_borrow_info, PathBound* const path_bound,
    std::string* const blocking_obstacle_id,
    std::string* const borrow_lane_type) {
  // 1.初始化边界.
  if (!InitPathBoundary(reference_line_info, path_bound)) {...}


  // 2.根据adc位置和lane信息确定大致的边界
  if (!GetBoundaryFromLanesAndADC(reference_line_info, lane_borrow_info, 0.1,
                                  path_bound, borrow_lane_type)) {...}
  // PathBoundsDebugString(*path_bound);

  // 3.根据障碍物调整道路边界
  if (!GetBoundaryFromStaticObstacles(reference_line_info.path_decision(),
                                      path_bound, blocking_obstacle_id)) {...}
  ...
}

流程和上面的几个基本类似,借道有三种类型文章来源地址https://www.toymoban.com/news/detail-672521.html

  enum class LaneBorrowInfo {
    LEFT_BORROW,
    NO_BORROW,
    RIGHT_BORROW,
  };

到了这里,关于【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER的文章就介绍完了。如果您还想了解更多内容,请在右上角搜索TOY模板网以前的文章或继续浏览下面的相关文章,希望大家以后多多支持TOY模板网!

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: 如若内容造成侵权/违法违规/事实不符,请点击违法举报进行投诉反馈,一经查实,立即删除!

领支付宝红包 赞助服务器费用

相关文章

  • 【Apollo学习笔记】——规划模块TASK之PIECEWISE_JERK_NONLINEAR_SPEED_OPTIMIZER(二)

    1.【Apollo学习笔记】——规划模块TASK之LANE_CHANGE_DECIDER 2.【Apollo学习笔记】——规划模块TASK之PATH_REUSE_DECIDER 3.【Apollo学习笔记】——规划模块TASK之PATH_BORROW_DECIDER 4.【Apollo学习笔记】——规划模块TASK之PATH_BOUNDS_DECIDER 5.【Apollo学习笔记】——规划模块TASK之PIECEWISE_JERK_PATH_OPTIMI

    2024年02月09日
    浏览(44)
  • Apollo规划模块代码学习(1): 算法架构原理、运行机制一文详解

    Apollo开源自动驾驶平台中,高清地图模块提供了每个在线模块都可以访问的高清地图。感知和定位模块提供了必要的动态环境信息,可以在预测模块中进一步用于预测未来的环境状态。运动规划模块考虑所有信息,以生成安全平滑的轨迹,并将其输入车辆控制模块。 目前Ap

    2024年01月25日
    浏览(47)
  • 【Apollo学习笔记】—— Routing模块

    Apollo的routing模块读取高精地图原始信息,用于根据输入 RoutingRequest 信息在 base_map 中选取匹配最近的点作为导航轨迹的起点和终点,读取依据 base_map 生成的 routing_map 作为 topo_graph ,然后通过Astar算法在拓扑图中搜索连接起始点的最优路径 RoutingResponse ,作为输出发送出去。

    2024年02月15日
    浏览(40)
  • Apollo星火计划学习笔记——Apollo开放空间规划算法原理与实践

    Apollo星火计划课程链接如下 星火计划2.0基础课:https://apollo.baidu.com/community/online-course/2 星火计划2.0专项课:https://apollo.baidu.com/community/online-course/12     开放空间算法的配置主要在 valet_parking_config.pb.txt 中,分为4个部分: OPEN_SPACE_ROI_DECIDER 、 OPEN_SPACE_TRAJECTORY_PROVIDER 、 OPE

    2023年04月10日
    浏览(39)
  • Apollo星火计划学习笔记——Apollo决策规划技术详解及实现(以交通灯场景检测为例)

    Apollo星火计划课程链接如下 星火计划2.0基础课:https://apollo.baidu.com/community/online-course/2 星火计划2.0专项课:https://apollo.baidu.com/community/online-course/12 星火计划学习笔记——第七讲自动驾驶规划技术原理1 ●目的 ○ 保障无人车的行车安全并且遵守交规 ○ 为路径和速度的平滑优

    2024年02月07日
    浏览(50)
  • Node.js 学习笔记 fs、path、http模块;模块化;包;npm

    教学视频 BV1MN411y7pw , P83-97的笔记 Node.js是一个跨平台JavaScript运行环境,使开发者可以搭建服务器端的JavaScript应用程序。 作用:使用 Node.js 编写服务器端程序 ✓ 编写数据接口,提供网页资源浏览功能等等 ✓ 前端工程化:为后续学习 Vue 和 React 等框架做铺垫 1.前端工程化 前

    2024年04月08日
    浏览(87)
  • 自动驾驶规划模块学习笔记-多项式曲线

    自动驾驶运动规划中会用到各种曲线,主要用于生成车辆变道的轨迹,高速场景中使用的是五次多项式曲线,城市场景中使用的是分段多项式曲线(piecewise),相比多项式,piecewise能够生成更为复杂的路径。另外对于自由空间,可以使用A*搜索出的轨迹再加以cilqr加以平滑,也

    2024年02月05日
    浏览(64)
  • Apollo Planning学习(2)-------路径规划

    本次学习的Apollo版本为6.0版本,因为从5.0开始轨迹规划算法主要使用的就是public road,所以本次主要学习该算法,该算法的核心思想是PV解耦,即Path-Velocity的解耦,其主要包含两个过程:1.路径规划,2.速度规划。 路径规划其实已经发展很多年,从早期的机器人到现在的无人驾

    2024年02月09日
    浏览(48)
  • Apollo Planning学习(9)-------速度规划

    顺着之前学习public road planner 的路径规划中lane follow的task,在得到的规划路径上再进行速度规划。大致思路为先利用ST Graph,将障碍物、限速等投影在ST图上,利用全局搜索方法DP算法得到决策,开辟一个凸空间,在利用最优化方法(二次优化和非线性优化)进行速度规划。 所

    2024年02月12日
    浏览(38)
  • Apollo决策规划算法学习Chapter3 速度规划算法

    第一章 Apollo决策规划算法基本概念 第二章 Apollo决策规划之路径规划算法 第三章 Apollo决策规划之速度规划算法 本文为第三章,主要讲解 Apollo决策规划算法中的速度规划算法,EM planner的速度规划算法同样是是通过动态规划和二次规划实现的,下面来细讲速度规划算法。 1)回

    2024年02月11日
    浏览(60)

觉得文章有用就打赏一下文章作者

支付宝扫一扫打赏

博客赞助

微信扫一扫打赏

请作者喝杯咖啡吧~博客赞助

支付宝扫一扫领取红包,优惠每天领

二维码1

领取红包

二维码2

领红包