Carla仿真学习笔记（版本0.9.16）

一、安装并编译 UE4

1 、 Carla 安装前的环境准备

（ubuntu20.04 ）：

|---------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| python3.8版本及以上；pip3正常使用；掉过头重新说明一下，由于官方carla0.9.16中的编译文件只有cp310、cp311、cp312（之前官方明确规定仅支持python3.7~3.9，此次carla也是最新版，可能有所改变），所以支持python3.10~3.12版本。本次选择python3.10.12 |

2 、 python 环境完成后，安装所需包

安装过程中如果遇到：

这里说明我们要更新 numpy 版本，然后按照提示安装 launchpadlib 后再安装：

|-------------------------------------------------------------------|
| Plain Text pip3 install --upgrade numpy pip3 install launchpadlib |

这里由于我们是使用 python 语言进行开发，所以这里最重要的就是安装 python3 ，其他环境安装如下：

3 、安装 Carla 和它的相关引擎

这里不同系统版本对应不同的安装指令【官网进行查看：How to build Carla on Ubuntu 】。这里使用的是 Ubuntu 20.04 版本，我们要保证Carla和它的引擎UnrealEngine用的是一个版本的Clang，命令如下：

4 、 UE4 安装

|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| 我们安装的是Carla0.9.16，从 Carla 0.9.12 版本开始，使用的是 Unreal Engine 4.26 的改进型分叉，此分叉包含 Carla 特有的补丁。在安装 Carla 0.9.16 时，需要先安装 Unreal Engine 4.26，可以通过将 GitHub 账户与 Unreal Engine 账户连接后，从 GitHub 上克隆 Carla 指定的 UE4 分支进行安装 |

a 、注册并登录账户

terminal安装git，并进行登录：

|---------------------------------------------------------------|
| Plain Text # 可在https://github.com/官网提前注册 sudo apt install git |

登录Feed - Unreal Engine，可以直接使用Epic账号邮箱密码登录

b 、接下来要完成 Github 和 UnrealEngine 的关联

连接方法可以参考该链接www.unrealengine.com，或者可以根据下面的方法：

使用经过验证的Epic Games帐户登录www.unrealengine.com，单击右上角网站标题栏中的用户名，然后选择personal选项。
从菜单中选择连接的账户（APPS AND ACCOUNTS），单击Github图标下方的连接。
查看《虚幻引擎最终用户许可协议》并选中对应复选框接受该协议，然后单击关联您的账户按钮（如果已经接受了最新版本的《虚幻引擎最终用户许可协议》，则不会显示该提示）。
登录现有GitHub帐户。
单击授权按钮，之后你会收到Gibhub发的一封EpicTeamAdmin邀请您加入确认的邮件，单击join @EpicGames完成关联流程。
需要进入https://github.com/EpicGames - Connect to preview 点击Follow

|----------------------------------------|
| 确认好上述均完成后，尤其是github和Rpic账号关联，再进行下面的操作。 |

c 、 Github Repo Clone

【这步主要是将UnrealEngine4.26git clone到自己ubuntu的home目录下，大概要花费90G---100G左右空间。】

|--------------------------------------------------------------------------------------------------------------|
| Plain Text git clone --depth 1 -b carla https://github.com/CarlaUnreal/UnrealEngine.git ~/UnrealEngine_4.26 |

可能会遇到git连接问题：

|--------------------------------------------------------------------------------------------------|
| 解决方法： 1.在项目文件夹的命令行窗口执行下面代码，然后再git commit 或git clone 取消git本身的https代理，使用自己本机的代理，如果没有的话，其实默认还是用git的 |

|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Plain Text //取消http代理 git config --global --unset http.proxy //取消https代理 git config --global --unset https.proxy // 要是还不行，试试 git config --global http.sslVerify "false" |

克隆过程中需要github的username（登录的用户名，不是邮箱 ）和password（GitHub 个人访问令牌（ Personal Access Token ））

以下是正常的界面：

5 、编译 UE4

|--------------------------------------------------------------------------------------------------------------------------|
| Plain Text cd UnrealEngine_4.26/ ./Setup.sh && ./GenerateProjectFiles.sh && make # 编译的时间可能很长.......因为，UE4.26分支源码规模很大 |

6 、测试 UE4 环境

|-----------------------------------------------------------------------------------|
| Plain Text # 启动UE4 cd ~/UnrealEngine_4.26/Engine/Binaries/Linux && ./UE4Editor |

|---------------------------------------------------------------------------|
| 测试完成，需要把UnrealEngine的路径添加到～/.bashrc中；末行添加"UE4_ROOT=~/UnrealEngine_4.26" |

二、安装 Carla0.9.16

找到相应的版本包：Release 0.9.16 · carla-simulator/carla（git下载有点慢）

下载完成后，解压（路径/home/trunk/carla_0.9.16/；其中carla目录需要自己提前创建）

进入carla_0.9.16/运行【./CarlaUE4.sh】，运行carla：

|-----------------------------------------------------------------|
| 目前成功安装，但运行结果不太满意：卡？硬件？着色器？------------------------显卡驱动未安装，调动显卡： |

三、测试 Carla ， Bug 解决中

|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Plain Text # 克隆官方的源码 git clone -b 0.9.16 https://github.com/carla-simulator/carla.git cd carla # ./Update.sh 的作用：只下载地图、车辆模型等资产文件; # Util/Setup.sh 的作用：安装编译必需的系统依赖（编译器、开发库等） ./Update.sh ./Setup.sh【版本不同Setup.sh所在位置不同】 |

1 、直接安装 Python3.10 及其以上版本

|----------------------------------------------------------------------------------|
| 因为carla0.9.16里面的whl文件只有cp301~cp312。之前尝试强制使用3.8版本去运行，失败；直接保留carla，更改python版本去适配。 |

2 、安装好之后，回到 dist 目录下，进行编译

3 、启动并测试

|--------------------------------------------------------------------------------------------------------------------------------------------------------|
| Plain Text # 启动引擎并加载 Town04 高速地图（后台运行，不占用终端） ./CarlaUE4.sh -carla-map=Town04 & # 引擎必须启动，才能执行后续脚本和方针运行 # 执行脚本测试 python3.10 manual_control.py 测试成功！ |

四、了解和学习命令

1 、了解文件和简单命令

|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Plain Text 执行manual_control.py（Carla****最基础的 "手动控制车辆" 示例 ）。键盘AWSD，分别控制左拐、前进加速、制动刹停、右拐；（S制动）空格刹车； P是切换自动驾驶和人工； Z是左转向灯，X是右转向灯（二者任一单独按两次即使双闪）； Q是切换前进/后退; C控制变换天气； |

a 、知悉脚本

automatic_control.py：纯自动驾驶脚本；
generate_traffic.py：批量生成自动驾驶的车辆和行人，自定义数量、车速、行为（如加塞、绕行），-n 车辆数、-p 行人数、--speed 最大车速；
vehicle_gallery.py：列出 Carla 所有可用车型
dynamic_weather.py：动态切换天气（晴、雨、雾、雪），还能调节时间（日出 / 日落）
start_recording.py ：记录仿真过程中的所有数据（车辆运动、传感器、交通灯、碰撞事件），保存为 .log 文件
start_replaying.py：回放之前记录的日志文件，重现场景（支持快进、慢放、指定时间点开始）
show_recorder_collisions.py：分析记录文件，显示所有碰撞事件（时间、碰撞的车辆 / 行人），用于事故分析
no_rendering_mode.py：关闭图形渲染（只保留物理仿真），大幅提升帧率，适合批量采集数据、训练算法（无需可视化）
synchronous_mode.py：同步模式（Python 脚本与 Carla 引擎帧率同步），避免数据丢失，适合高精度仿真
tutorial.py：Carla Python API 入门教程，一步步教你创建车辆、添加传感器、控制运动，带详细注释。
V2XDemo.py：车联网（V2X）演示，车辆之间、车辆与路侧设备通信，适合车路协同场景测试。
自定义脚本：完全自己写的，可以尝试一下··················

b 、地图选择

|-------------------------------------------------|
| Plain Text # 按"~"进入控制台，然后使用命令"Open Town04"切换地图 |

|-----------------------------------------------------------------------------------------------------------------------------------------|
| Plain Text # 目前只有Town01~05的五张基础地图，其余需要可以下载拓展地图（官网） # 下载官方拓展地图包，然后放到carla_0.9.16/Import/下，然后执行ImportAssets.sh脚本即可。 # 下面是Town06,模拟的高速道路。 |

2 、实现简单的交互场景

源代码：

|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Python import carla import random import time import subprocess import os # 核心配置 CARLA_PATH = "/home/trunk/carla_0.9.16" HOST = '127.0.0.1' PORT = 2000 MAP_NAME = "Town06" EGO_VEHICLE_COLOR = "255,0,0" # 自车红色 EGO_SPEED_KMH = 85 STATIC_CARS_NUM = 3 LEFT_CARS_NUM = 8 RIGHT_CARS_NUM = 8 MIDDLE_MOVING_CARS = 2 SIDE_CARS_SPEED_MIN = 75 SIDE_CARS_SPEED_MAX = 82 FIXED_VIEW_LOCATION = carla.Location(x=100, y=0, z=35) FIXED_VIEW_ROTATION = carla.Rotation(pitch=-70, yaw=0) def start_carla_server(): """检查并启动 Carla 服务器""" print("正在检查 Carla 服务器是否运行...") try: client = carla.Client(HOST, PORT) client.set_timeout(2.0) client.get_world() print("Carla 服务器已在运行。") return True except RuntimeError: print("未检测到 Carla 服务器，正在尝试启动...") carla_sh_path = os.path.join(CARLA_PATH, "CarlaUE4.sh") if not os.path.exists(carla_sh_path): print(f"错误：无法找到 Carla 启动脚本 {carla_sh_path}") return False subprocess.Popen([carla_sh_path, f"-carla-map={MAP_NAME}"], stdout=subprocess.PIPE, stderr=subprocess.PIPE) print("等待 Carla 服务器启动... (30-60 秒)") time.sleep(30) try: client = carla.Client(HOST, PORT) client.set_timeout(10.0) client.get_world() print("Carla 服务器启动成功！") return True except RuntimeError as e: print(f"错误：启动 Carla 服务器失败。{e}") return False def is_straight_waypoint(waypoint, distance=15.0): """判断路点是否为直线（适配 Carla 0.9.16）""" next_waypoints = waypoint.next(distance) if not next_waypoints: return False next_waypoint = next_waypoints[0] current_yaw = waypoint.transform.rotation.yaw next_yaw = next_waypoint.transform.rotation.yaw yaw_diff = abs(current_yaw - next_yaw) yaw_diff = min(yaw_diff, 360 - yaw_diff) return yaw_diff < 1.5 def get_three_consecutive_lanes(world): """筛选连续三条直线车道""" spawn_points = world.get_map().get_spawn_points() straight_spawns = [] for sp in spawn_points: waypoint = world.get_map().get_waypoint(sp.location) if waypoint.lane_type == carla.LaneType.Driving and is_straight_waypoint(waypoint): straight_spawns.append((sp, waypoint.lane_id)) lane_spawns = {} for sp, lane_id in straight_spawns: if lane_id not in lane_spawns: lane_spawns[lane_id] = [] lane_spawns[lane_id].append(sp) sorted_lane_ids = sorted(lane_spawns.keys()) total_lanes = len(sorted_lane_ids) if total_lanes < 3: raise Exception(f"当前地图仅找到 {total_lanes} 条直线车道，不足3条无法测试") best_consecutive = None max_spawns = 0 for i in range(total_lanes - 2): consecutive_ids = sorted_lane_ids[i:i+3] total_sp = sum(len(lane_spawns[lid]) for lid in consecutive_ids) if total_sp > max_spawns: max_spawns = total_sp best_consecutive = consecutive_ids left_lane_id = best_consecutive[0] middle_lane_id = best_consecutive[1] right_lane_id = best_consecutive[2] print(f"已筛选连续三条直线车道：左车道{left_lane_id} → 中间车道{middle_lane_id} → 右车道{right_lane_id}") return { "left": lane_spawns[left_lane_id], "middle": lane_spawns[middle_lane_id], "right": lane_spawns[right_lane_id] } def get_black_car_bp(world): """获取黑色小汽车蓝图（排除两轮车，确保有颜色属性）""" # 筛选：四轮汽车 + 排除两轮车 + 有颜色属性 valid_cars = [] for bp in world.get_blueprint_library().filter('vehicle.*'): # 排除两轮车 if any(keyword in bp.id for keyword in ['motorcycle', 'bike', 'bicycle', 'scooter']): continue # 确保有颜色属性 if bp.has_attribute('color'): valid_cars.append(bp) if not valid_cars: raise Exception("错误：未找到符合条件的小汽车（需支持颜色设置且非两轮车）") # 优先选紧凑型/小型车 compact_cars = [bp for bp in valid_cars if bp.id.startswith('vehicle.compact') or bp.id.startswith('vehicle.micro')] bp = random.choice(compact_cars) if compact_cars else random.choice(valid_cars) bp.set_attribute('color', '0,0,0') # 强制黑色 return bp def get_red_truck_bp(world): """获取红色带挂货车蓝图（确保有颜色属性）""" # 筛选：货车/挂车 + 有颜色属性 valid_trucks = [] for bp in world.get_blueprint_library().filter('vehicle.*'): # 匹配货车/挂车模型（包含truck、semi关键词） if 'truck' in bp.id or 'semi' in bp.id: if bp.has_attribute('color'): valid_trucks.append(bp) if not valid_trucks: raise Exception("错误：未找到符合条件的货车（需支持颜色设置）") # 优先选带挂的货车（semi类） semi_trucks = [bp for bp in valid_trucks if 'semi' in bp.id] bp = random.choice(semi_trucks) if semi_trucks else random.choice(valid_trucks) bp.set_attribute('color', EGO_VEHICLE_COLOR) # 强制红色 return bp def main(): if not start_carla_server(): return try: client = carla.Client(HOST, PORT) client.set_timeout(15.0) world = client.load_world(MAP_NAME) print(f"地图 {MAP_NAME} 加载成功") # 调整为正常天气 weather = carla.WeatherParameters( cloudiness=40.0, precipitation=0.0, sun_altitude_angle=45.0, fog_density=0.03 ) world.set_weather(weather) print("已设置为正常天气（柔和阴天）") tm = client.get_trafficmanager(8000) tm.set_global_distance_to_leading_vehicle(2.8) tm.set_hybrid_physics_mode(True) tm.set_hybrid_physics_radius(500.0) print("正在筛选连续三条直线车道...") lane_spawns = get_three_consecutive_lanes(world) for lane, spawns in lane_spawns.items(): print(f"{lane}车道可用生成点：{len(spawns)} 个") # 1. 中间车道：生成3辆静止车（黑色小汽车） print(f"\n生成中间车道静止车辆（{STATIC_CARS_NUM} 辆，黑色小汽车）...") middle_spawns = lane_spawns["middle"] static_cars = [] for i in range(STATIC_CARS_NUM): base_spawn = middle_spawns[0] static_location = carla.Location( x=base_spawn.location.x + 90 + i*15, y=base_spawn.location.y, z=base_spawn.location.z ) static_transform = carla.Transform(static_location, base_spawn.rotation) bp = get_black_car_bp(world) try: car = world.spawn_actor(bp, static_transform) car.set_autopilot(False) static_cars.append(car) print(f"生成静止车 {i+1}: {bp.id}（位置：x={static_location.x:.1f}）") except Exception as e: print(f"生成静止车 {i+1} 失败: {e}") # 2. 中间车道：生成自车（红色带挂货车）+ 2辆额外行驶车（黑色小汽车） print("\n生成中间车道车辆...") # 自车（红色带挂货车） base_spawn = middle_spawns[0] ego_spawn_location = carla.Location( x=base_spawn.location.x + 20, y=base_spawn.location.y, z=base_spawn.location.z ) ego_transform = carla.Transform(ego_spawn_location, base_spawn.rotation) ego_bp = get_red_truck_bp(world) ego_vehicle = world.spawn_actor(ego_bp, ego_transform) ego_vehicle.set_autopilot(True, tm.get_port()) ego_speed_factor = EGO_SPEED_KMH / 120.0 tm.vehicle_percentage_speed_difference(ego_vehicle, int((ego_speed_factor - 1.0) * 100)) print(f"生成自车: {ego_bp.id}（红色带挂货车，速度：{EGO_SPEED_KMH} km/h，位置：x={ego_spawn_location.x:.1f}）") # 中间车道额外2辆行驶车（黑色小汽车） middle_moving_cars = [] for i in range(MIDDLE_MOVING_CARS): if i == 0: move_location = carla.Location( x=base_spawn.location.x + 50, y=base_spawn.location.y, z=base_spawn.location.z ) speed = random.randint(EGO_SPEED_KMH-5, EGO_SPEED_KMH) else: move_location = carla.Location( x=base_spawn.location.x - 20, y=base_spawn.location.y, z=base_spawn.location.z ) speed = random.randint(EGO_SPEED_KMH-8, EGO_SPEED_KMH-3) move_transform = carla.Transform(move_location, base_spawn.rotation) bp = get_black_car_bp(world) try: car = world.spawn_actor(bp, move_transform) car.set_autopilot(True, tm.get_port()) speed_factor = speed / 120.0 tm.vehicle_percentage_speed_difference(car, int((speed_factor - 1.0) * 100)) middle_moving_cars.append(car) print(f"生成中间车道行驶车 {i+1}: {bp.id}（黑色小汽车，速度：{speed} km/h，位置：x={move_location.x:.1f}）") except Exception as e: print(f"生成中间车道行驶车 {i+1} 失败: {e}") # 3. 左侧车道：生成8辆车辆（黑色小汽车） print(f"\n生成左侧车道车辆（{LEFT_CARS_NUM} 辆，黑色小汽车）...") left_spawns = lane_spawns["left"] left_cars = [] for i in range(LEFT_CARS_NUM): spawn_point = left_spawns[i % len(left_spawns)] spawn_location = carla.Location( x=spawn_point.location.x + random.randint(-80, 80), y=spawn_point.location.y, z=spawn_point.location.z ) spawn_transform = carla.Transform(spawn_location, spawn_point.rotation) bp = get_black_car_bp(world) try: car = world.spawn_actor(bp, spawn_transform) car.set_autopilot(True, tm.get_port()) random_speed = random.randint(SIDE_CARS_SPEED_MIN, SIDE_CARS_SPEED_MAX) speed_factor = random_speed / 120.0 tm.vehicle_percentage_speed_difference(car, int((speed_factor - 1.0) * 100)) left_cars.append(car) if (i+1) % 4 == 0: print(f"生成左侧车 {i+1}: {bp.id}（速度：{random_speed} km/h，位置：x={spawn_location.x:.1f}）") except Exception as e: print(f"生成左侧车 {i+1} 失败: {e}") # 4. 右侧车道：生成8辆车辆（黑色小汽车） print(f"\n生成右侧车道车辆（{RIGHT_CARS_NUM} 辆，黑色小汽车）...") right_spawns = lane_spawns["right"] right_cars = [] for i in range(RIGHT_CARS_NUM): spawn_point = right_spawns[i % len(right_spawns)] spawn_location = carla.Location( x=spawn_point.location.x + random.randint(-80, 80), y=spawn_point.location.y, z=spawn_point.location.z ) spawn_transform = carla.Transform(spawn_location, spawn_point.rotation) bp = get_black_car_bp(world) try: car = world.spawn_actor(bp, spawn_transform) car.set_autopilot(True, tm.get_port()) random_speed = random.randint(SIDE_CARS_SPEED_MIN, SIDE_CARS_SPEED_MAX) speed_factor = random_speed / 120.0 tm.vehicle_percentage_speed_difference(car, int((speed_factor - 1.0) * 100)) right_cars.append(car) if (i+1) % 4 == 0: print(f"生成右侧车 {i+1}: {bp.id}（速度：{random_speed} km/h，位置：x={spawn_location.x:.1f}）") except Exception as e: print(f"生成右侧车 {i+1} 失败: {e}") # 5. 固定高空俯览视角 print("\n设置固定视角...") spectator = world.get_spectator() spectator.set_transform(carla.Transform(FIXED_VIEW_LOCATION, FIXED_VIEW_ROTATION)) print("视角已固定（高空俯览，覆盖三条测试车道）") # 场景就绪提示 total_cars = 1 + STATIC_CARS_NUM + MIDDLE_MOVING_CARS + LEFT_CARS_NUM + RIGHT_CARS_NUM print("\n" + "-"*50) print("变道测试场景已就绪（Carla 0.9.16）") print(f"总车辆数：{total_cars} 辆") print(f"自车：红色带挂货车（{ego_bp.id}，中间车道，{EGO_SPEED_KMH} km/h）") print("其他车辆：全部为黑色小汽车（无两轮车）") print("天气：正常柔和阴天") print("观察重点：红色货车接近静止车时是否主动变道避让") print("-"*50) print("按 Ctrl+C 终止脚本并清理车辆") while True: time.sleep(1) except KeyboardInterrupt: print("\n正在清理所有车辆...") actors = world.get_actors() vehicles = actors.filter('vehicle.*') vehicle_count = len(vehicles) for v in vehicles: v.destroy() print(f"已销毁所有 {vehicle_count} 辆车辆") except Exception as e: print(f"\n场景运行出错: {e}") actors = world.get_actors() vehicles = actors.filter('vehicle.*') for v in vehicles: v.destroy() print("已自动清理所有车辆") if name == 'main': main() |

图示：

3 、设置模拟目标点场景搭建

|-----------------------------------------------------------------------|
| 这里采用：对每个车辆（包括自车）设置初始和目标位置（包含多个目标点），设置初始和目标点的速度大小。（后续在此脚本上进行更改和变化即可） |

|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Python import carla import time import math # --- 配置参数 --- HOST = '127.0.0.1' PORT = 2000 # --- 核心配置 --- # 每辆车都包含：初始位置、初始速度、以及一系列目标点（位置+速度） VEHICLE_CONFIGS = { 'ego': { # 自车 (消防车车型) 'model': 'vehicle.carlamotors.firetruck', 'initial_transform': carla.Transform( carla.Location(x=22, y=143, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, # 初始速度为0 'target_points': [ {'location': carla.Location(x=222, y=143, z=0.3), 'speed': 80.0}, {'location': carla.Location(x=322, y=143, z=0.3), 'speed': 60.0}, {'location': carla.Location(x=400, y=146.5, z=0.3), 'speed': 60.0}, {'location': carla.Location(x=622, y=143, z=0.3), 'speed': 0.0} ] }, # 右侧车辆 'r_0': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=41, y=146.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=241, y=146.5, z=0.3), 'speed': 60.0}, {'location': carla.Location(x=441, y=146.5, z=0.3), 'speed': 0.0}, ] }, 'r_1': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=34, y=146.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=234, y=146.5, z=0.3), 'speed': 60.0}, {'location': carla.Location(x=434, y=146.5, z=0.3), 'speed': 0.0}, ] }, 'r_2': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=27, y=146.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=227, y=142.587830, z=0.3), 'speed': 0.0}, {'location': carla.Location(x=327, y=142.587830, z=0.3), 'speed': 00.0}, ] }, 'r_3': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=20, y=146.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=220, y=142.587830, z=0.3), 'speed': 0.0}, {'location': carla.Location(x=320, y=142.587830, z=0.3), 'speed': 00.0}, ] }, # 左侧车辆 'l_0': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=41, y=139.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=241, y=139.5, z=0.3), 'speed': 60.0}, {'location': carla.Location(x=441, y=139.5, z=0.3), 'speed': 0.0}, ] }, 'l_1': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=34, y=139.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=234, y=139.5, z=0.3), 'speed': 60.0}, {'location': carla.Location(x=434, y=139.5, z=0.3), 'speed': 00.0}, ] }, 'l_2': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=27, y=139.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=227, y=142.587830, z=0.3), 'speed': 0.0}, {'location': carla.Location(x=327, y=142.587830, z=0.3), 'speed': 00.0}, ] }, 'l_3': { # 黑色小汽车 'model': 'vehicle.tesla.model3', 'initial_transform': carla.Transform( carla.Location(x=20, y=139.5, z=0.300000), carla.Rotation(pitch=0, yaw=0, roll=0) ), 'initial_speed': 0.0, 'target_points': [ {'location': carla.Location(x=220, y=142.587830, z=0.3), 'speed': 0.0}, {'location': carla.Location(x=320, y=142.587830, z=0.3), 'speed': 00.0}, ] } # 可以继续添加更多车辆... } # --- 辅助函数 --- def clear_existing_actors(world): """清除世界中所有现有车辆""" print("正在清除现有车辆...") for actor in world.get_actors().filter('vehicle.*'): actor.destroy() print("现有车辆已清除。") def generate_path(world, start_location, end_location, resolution=2.0): """生成从起点到终点的路径点列表""" path = [] start_waypoint = world.get_map().get_waypoint(start_location) end_waypoint = world.get_map().get_waypoint(end_location) # 如果起点和终点在同一条车道上，直接生成直线路径 if start_waypoint.lane_id == end_waypoint.lane_id: current_waypoint = start_waypoint while current_waypoint.transform.location.distance(end_location) > resolution: path.append(current_waypoint.transform.location) next_waypoints = current_waypoint.next(resolution) if not next_waypoints: break current_waypoint = next_waypoints[0] path.append(end_location) else: # 这里简化处理，实际复杂场景可能需要更高级的路径规划 path.append(start_location) path.append(end_location) return path def control_vehicle(vehicle, target_location, target_speed): """ 控制车辆朝向目标位置行驶，并尝试达到目标速度返回: 车辆是否已到达目标位置 """ vehicle_location = vehicle.get_transform().location distance = vehicle_location.distance(target_location) # 如果距离目标点很近，认为已经到达 if distance < 3.0: return True # 计算朝向目标点所需的转向角 vehicle_rotation = vehicle.get_transform().rotation forward_vector = carla.Vector3D( math.cos(math.radians(vehicle_rotation.yaw)), math.sin(math.radians(vehicle_rotation.yaw)), 0 ) target_vector = carla.Vector3D( target_location.x - vehicle_location.x, target_location.y - vehicle_location.y, 0 ) target_vector = target_vector.make_unit_vector() # 计算两个向量的夹角 dot_product = forward_vector.x * target_vector.x + forward_vector.y * target_vector.y cross_product = forward_vector.x * target_vector.y - forward_vector.y * target_vector.x angle = math.degrees(math.atan2(cross_product, dot_product)) # 速度控制 current_speed = vehicle.get_velocity().length() speed_error = (target_speed / 3.6) - current_speed throttle = 0.0 brake = 0.0 steer = angle / 45.0 # 将角度归一化到[-1, 1] if speed_error > 0: throttle = min(1.0, speed_error / 5.0) else: brake = min(1.0, -speed_error / 2.0) # 应用控制 vehicle.apply_control(carla.VehicleControl(throttle=throttle, steer=steer, brake=brake)) return False # --- 主控制循环 --- def main(): try: client = carla.Client(HOST, PORT) client.set_timeout(10.0) world = client.get_world() world.set_weather(carla.WeatherParameters.ClearNoon) clear_existing_actors(world) # 生成车辆并存储其状态 vehicles = {} blueprint_library = world.get_blueprint_library() for vehicle_id, config in VEHICLE_CONFIGS.items(): blueprint = blueprint_library.find(config['model']) if not blueprint: print(f"警告: 未找到车型 {config['model']}，跳过。") continue if config['model'] != 'vehicle.carlamotors.firetruck': blueprint.set_attribute('color', '0,0,0') vehicle = world.spawn_actor(blueprint, config['initial_transform']) if not vehicle: print(f"警告: 生成车辆 {vehicle_id} 失败。") continue vehicle.set_simulate_physics(True) # 初始速度为0 vehicle.apply_control(carla.VehicleControl(throttle=0.0, brake=1.0)) time.sleep(0.1) vehicle.apply_control(carla.VehicleControl(throttle=0.0, brake=0.0)) vehicles[vehicle_id] = { 'actor': vehicle, 'target_points': config['target_points'], 'current_target_index': 0, 'current_path': [] } print(f"成功生成车辆 ID: {vehicle_id}") print("\n开始车辆控制...") while True: for vehicle_id, state in vehicles.items(): vehicle = state['actor'] target_points = state['target_points'] current_index = state['current_target_index'] if current_index >= len(target_points): continue # 所有目标点都已完成 current_target = target_points[current_index] # 如果当前没有路径，生成一条 if not state['current_path']: print(f"车辆 {vehicle_id}：生成到目标点 {current_index} 的路径...") state['current_path'] = generate_path( world, vehicle.get_transform().location, current_target['location'] ) if not state['current_path']: print(f"警告: 车辆 {vehicle_id} 无法生成到目标点 {current_index} 的路径。") state['current_target_index'] += 1 continue # 获取下一个路径点 next_waypoint_location = state['current_path'][0] # 控制车辆向路径点移动 reached_waypoint = control_vehicle( vehicle, next_waypoint_location, current_target['speed'] ) # 如果到达了路径点，将其从列表中移除 if reached_waypoint: state['current_path'].pop(0) # 如果路径为空，说明到达了目标点 if not state['current_path']: print(f"车辆 {vehicle_id}：已到达目标点 {current_index}") state['current_target_index'] += 1 if state['current_target_index'] >= len(target_points): print(f"车辆 {vehicle_id}：所有目标点已完成！") time.sleep(0.05) except KeyboardInterrupt: print("\n程序被手动停止。") finally: print("\n正在清理资源...") for vehicle_id, state in vehicles.items(): try: state['actor'].destroy() print(f"已销毁车辆 ID: {vehicle_id}") except Exception as e: print(f"销毁车辆 ID: {vehicle_id} 时出错: {e}") print("所有资源已清理。") if name == 'main': main() |

五、（预先拓展）结合 ROS 环境，桥接

1. 安装 ROS Noetic （已经存在，验证）

2. 配置 CARLA-ROS 桥接

2.1 创建并进入 Catkin 工作空间

|---------------------------------------------------------------------|
| Plain Text # 创建工作空间目录 mkdir -p ~/catkin_ws/src cd ~/catkin_ws/src |

2.2 下载兼容 CARLA 0.9.16 的 ROS 桥接包

2.3 安装桥接包依赖

2.4 编译工作空间

2.5 同时添加 CARLA 环境变量（避免每次编译 / 运行脚本都要手动设置）：

3. 开始连接并进行测试（有问题： ros Noetic 是依赖 python3.8 ，在原 Ubuntu 系统中已经存在）

4. 创建 Python3.10 的虚拟环境

5. 桥接步骤（有些修改放在步骤 7）

6. ROS和 Carla 桥接成功！！！！！！！！

7. 备注信息（消除版本影响和取消地图选择【即默认地图选择】、修复关闭 bug ）

|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Python **#**在" ~/catkin_ws/src/carla_ros_bridge/carla_ros_bridge/src/carla_ros_bridge/bridge.py" 找到" def destroy(self):" # self.shutdown.set() # 注释掉或删除这行 self.shutdown_event.set() |

8 、重新安装 carla_ws 工作环境（即桥接环境）

|------------------------------------------------------------------------------------------------------------------------------------------------|
| catkin_make 和 catkin build 是两种不同的 ROS 编译工具，设计理念和规则差异很大 ------ 之前用 catkin_make 不行，换 catkin build 就成了，本质是工具的 "规则适配性" 问题。 |

Carla 0.9.16 + ROS Noetic + Python3.10 + 0.9.13****桥接包 =====>>> 成功链接

六、桥接测试，利用 ROS 生态和 Carla 闭环联动

1. 利用 ros 查看和播放 bag 包

|----------------------------------------|
| Python # bag包使用的是"ID25059" rosbag info |

2. ros_bridge 和 carla 链接操作步骤

|--------------------------------------------------------------------------------------------------------|
| 流程：启动 Carla → 手动生成 Ego 车 → 启动 ROS Bridge 和转换节点 → 回放 bag 包 |

|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Python # 正常启动ros-carla连接 # 先启动carla（窗口化） ./CarlaUE4.sh -windowed -ResX=1280 -ResY=720 # 再启动roslaunch文件（记得使用python虚拟环境） source devel/setup.sh roslaunch carla_ros_bridge carla_ros_bridge.launch |

3 、创建 pncc_msgs 包，驱动 ego 车辆

3.1 创建自定义消息包

3.2 测试 pncc_msgs 和 ros 正常对接

|------------------------------------------------------------------------------------------|
| 单纯的 rostopic pub/echo 测试确实是基础验证步骤，它的核心价值是【排除 pncc_msgs 本身的问题】，为后续【bag 转译 + Carla 车辆复现】铺路 |