diff --git a/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/car_env.py b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/car_env.py index 230a4e6cf1..c9d79f4a81 100644 --- a/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/car_env.py +++ b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/car_env.py @@ -4,6 +4,7 @@ import os import sys import random +import config as cfg try: sys.path.append(glob.glob('../carla/dist/carla-*%d.%d-%s.egg' % ( @@ -13,12 +14,13 @@ except IndexError: pass -# 获取当前文件的目录 +# 获取当前文件目录 current_dir = os.path.dirname(os.path.abspath(__file__)) -# 获取父目录 -parent_dir = os.path.dirname(current_dir) -# 将父目录添加到系统路径 -sys.path.append(parent_dir) +# 构建可能的路径 +possible_paths = os.path.join(current_dir, '..', 'carla_ros_ws', 'src','carla_autonomous') +# 找到第一个存在的路径并添加 +abs_path = os.path.abspath(possible_paths) +sys.path.insert(0, abs_path) from agents.navigation.global_route_planner import GlobalRoutePlanner import carla @@ -413,6 +415,7 @@ def process_images(self): self.distance = min(dis, dis_ped) + self.actual_distance = min(dis, dis_ped) # 保存实际距离 return dis @@ -486,7 +489,16 @@ def step(self, action, current_state): signed_dis = 0 current_state[3] = current_state[3]/15 - + + if hasattr(self, 'actual_distance'): + actual_obstacle_distance = self.actual_distance + else: + actual_obstacle_distance = 0 + + print(f"Obstacle Distance: {actual_obstacle_distance:.2f}m") + print(f"Angle Error: {phi:.2f}°, Lateral Error: {signed_dis:.2f}m") + print(f"Speed: {kmh:.2f}km/h, Action: {cfg.ACTION_NAMES[action]}") + print(f"角度差: {phi:.2f}°, 横向偏差: {signed_dis:.2f}m") print(f"距离障碍物: {current_state[0]*300+300:.2f}m") print(f"速度: {(current_state[1]+current_state[0])*30+30:.2f}km/h") @@ -515,15 +527,12 @@ def step(self, action, current_state): reward = 1000 print("✅ 成功到达终点!") - # 超时检查 - if self.episode_start + 200 < time.time(): - done = True - print("⏰ 超时") - print(f"奖励: {reward}") # 返回新的状态、奖励、完成标志和当前路径点 - return [(self.distance-300)/300, (kmh-30)/30-(self.distance-300)/300, phi, signed_dis*15], reward, done, current_waypoint + norm_distance = (self.actual_distance - 300) / 300 if hasattr(self, 'actual_distance') else 0 + return [norm_distance, (kmh-30)/30-norm_distance, phi, signed_dis*15], reward, done, current_waypoint + def trajectory(self, draw = False): amap = self.world.get_map() diff --git a/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/config.py b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/config.py index cd89a466af..7a5d66b7f1 100644 --- a/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/config.py +++ b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/config.py @@ -6,7 +6,7 @@ TRAJECTORIES = { "custom_trajectory": { "start": [108.91605377197266,101.3561782836914,0.6136533617973328, 0], - "end": [74.64781951904297,-56.014076232910156,0.40768954157829285], + "end": [62.20726013183594,-68.12548065185547,0.7430136203765869], "description": "自定义轨迹 - 城镇道路" }, "test_trajectory": { @@ -32,7 +32,7 @@ def get_current_trajectory(): # 设置基础路径 current_dir = os.path.dirname(os.path.abspath(__file__)) base_dir = os.path.dirname(current_dir) # 上一级目录 -MODELS_DIR = os.path.join(base_dir, 'models') +MODELS_DIR = os.path.join(base_dir, 'carla_ros_ws/src/carla_autonomous/models') braking = os.path.join(MODELS_DIR, 'Braking___282.model') driving = os.path.join(MODELS_DIR, 'Driving__6030.model') MODEL_PATHS = { @@ -126,6 +126,21 @@ def get_current_trajectory(): CAMERA_WIDTH = 640 CAMERA_HEIGHT = 480 CAMERA_FOV = 40 +# ==================== 数据收集配置 ==================== +ENABLE_DATA_COLLECTION = True # 是否启用数据收集 +DATA_SAVE_DIR = "data_logs" # 数据保存目录 +ENABLE_PLOTTING = True # 是否生成图表 +PLOT_DPI = 150 # 图表DPI +AUTO_OPEN_PLOTS = False # 是否自动打开图表(可能不支持无GUI环境) + +# 数据采样率(1表示记录每一步,2表示每两步记录一次等) +DATA_SAMPLING_RATE = 1 + + +# ==================== 图表字体配置 ==================== +CHART_FONT_FAMILY = 'SimHei' # 默认字体 +CHART_FONT_SIZE = 12 +ENABLE_CHINESE_FONT = True def print_config(): """打印当前配置""" @@ -161,4 +176,10 @@ def print_config(): print(f" 固定时间步长: {FIXED_DELTA_SECONDS}s") print(f" 同步模式: {'开启' if SYNCHRONOUS_MODE else '关闭'}") + print("数据收集配置\n") + print(f"数据收集: {'启用' if ENABLE_DATA_COLLECTION else '禁用'}") + print(f"图表生成: {'启用' if ENABLE_PLOTTING else '禁用'}") + print(f"数据保存目录: {DATA_SAVE_DIR}") + print(f"数据采样率: 每{DATA_SAMPLING_RATE}步记录一次") + print("="*60) \ No newline at end of file diff --git a/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/data_collector.py b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/data_collector.py new file mode 100644 index 0000000000..a73b1d1d38 --- /dev/null +++ b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/data_collector.py @@ -0,0 +1,609 @@ +""" +数据收集和可视化模块 - 收集运行数据并绘制图表 +使用中文标注,确保图表正确显示中文 +""" + +import time +import numpy as np +import matplotlib.pyplot as plt +from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas +import matplotlib +matplotlib.use('Agg') # 使用非GUI后端 + +# 设置matplotlib使用中文字体 +import matplotlib.font_manager as fm +import os + +# 创建一个专门的中文字体设置函数 +def setup_chinese_font(): + """设置中文字体,兼容Windows、Linux、Mac系统""" + + # 常见中文字体文件路径 + font_paths = [ + # Windows + "C:/Windows/Fonts/simhei.ttf", # 黑体 + "C:/Windows/Fonts/simsun.ttc", # 宋体 + "C:/Windows/Fonts/simkai.ttf", # 楷体 + "C:/Windows/Fonts/simfang.ttf", # 仿宋 + "C:/Windows/Fonts/msyh.ttc", # 微软雅黑 + "C:/Windows/Fonts/msyhbd.ttc", # 微软雅黑粗体 + + # Linux + "/usr/share/fonts/truetype/wqy/wqy-microhei.ttc", # 文泉驿微米黑 + "/usr/share/fonts/truetype/wqy/wqy-zenhei.ttc", # 文泉驿正黑 + "/usr/share/fonts/truetype/arphic/uming.ttc", # 文鼎明体 + + # Mac + "/System/Library/Fonts/STHeiti Medium.ttc", # 黑体 + "/System/Library/Fonts/STSong.ttf", # 宋体 + "/System/Library/Fonts/AppleGothic.ttf", # 苹果字体 + + # 通用 + "simhei.ttf", + "msyh.ttc", + ] + + # 先尝试直接设置字体名称(适用于已安装字体的系统) + font_names = [ + 'SimHei', # Windows黑体 + 'Microsoft YaHei', # Windows微软雅黑 + 'STHeiti', # Mac黑体 + 'STSong', # Mac宋体 + 'WenQuanYi Micro Hei', # Linux文泉驿微米黑 + 'DejaVu Sans', # 回退字体 + 'Arial', # 基本字体 + ] + + # 尝试使用字体名称 + for font_name in font_names: + try: + # 检查字体是否可用 + if any(font_name.lower() in f.name.lower() for f in fm.fontManager.ttflist): + plt.rcParams['font.sans-serif'] = [font_name] + plt.rcParams['axes.unicode_minus'] = False + print(f"✅ 使用系统字体: {font_name}") + return True + except: + continue + + # 如果系统字体不可用,尝试从文件加载 + for font_path in font_paths: + if os.path.exists(font_path): + try: + # 添加字体 + fm.fontManager.addfont(font_path) + font_prop = fm.FontProperties(fname=font_path) + font_name = font_prop.get_name() + + plt.rcParams['font.sans-serif'] = [font_name] + plt.rcParams['axes.unicode_minus'] = False + print(f"✅ 加载字体文件: {font_name} ({font_path})") + return True + except Exception as e: + print(f"⚠️ 加载字体失败 {font_path}: {e}") + continue + + # 如果所有方法都失败,尝试生成一个临时的中文字体解决方案 + print("⚠️ 无法找到合适的中文字体,将尝试使用回退方案") + + # 设置默认字体 + plt.rcParams['font.sans-serif'] = ['DejaVu Sans', 'Arial'] + plt.rcParams['axes.unicode_minus'] = False + + return False + +# 设置中文字体 +setup_chinese_font() + +import pandas as pd +from datetime import datetime +import config as cfg + +class DataCollector: + """数据收集和可视化类""" + + def __init__(self, save_dir="data_logs"): + self.save_dir = save_dir + self.start_time = None + self.end_time = None + self.episode_start_time = None + + # 检查字体是否设置成功 + self.chinese_font_available = self._check_chinese_font() + + # 创建保存目录 + if not os.path.exists(save_dir): + os.makedirs(save_dir) + + # 创建绘图目录 + self.plot_dir = os.path.join(save_dir, "plots") + if not os.path.exists(self.plot_dir): + os.makedirs(self.plot_dir) + + # 数据存储列表 + self.timestamps = [] # 时间戳(相对运行时间) + self.real_timestamps = [] # 实际时间戳 + self.obstacle_distances = [] # 障碍物距离 + self.actions = [] # 执行的动作 + self.lateral_errors = [] # 横向偏差 + self.angle_errors = [] # 角度差 + self.speeds = [] # 速度 + self.fps_values = [] # 帧率 + self.rewards = [] # 奖励值 + + # 用于计算帧率的变量 + self.last_time = None + self.frame_count = 0 + self.frame_times = [] + + # 动作名称映射(中文) + self.action_names_cn = ["刹车", "直行", "左转", "右转", "微左", "微右"] + + print(f"📊 数据收集器初始化完成,数据将保存到: {save_dir}") + print(f"📝 中文字体可用: {'是' if self.chinese_font_available else '否'}") + + def _check_chinese_font(self): + """检查中文字体是否可用""" + try: + # 尝试绘制一个包含中文的简单图形来测试字体 + fig, ax = plt.subplots(figsize=(1, 1)) + ax.text(0.5, 0.5, "测试", fontsize=12) + plt.close(fig) + return True + except: + return False + + def start_episode(self): + """开始一个episode的数据收集""" + self.episode_start_time = time.time() + self.timestamps = [] + self.real_timestamps = [] + self.obstacle_distances = [] + self.actions = [] + self.lateral_errors = [] + self.angle_errors = [] + self.speeds = [] + self.fps_values = [] + self.rewards = [] + self.last_time = time.time() + self.frame_count = 0 + self.frame_times = [] + + print("🔄 开始记录episode数据") + + def record_step(self, env, action, current_state, reward, vehicle_state=None): + """记录每一步的数据""" + current_time = time.time() + + # 计算相对时间 + if self.episode_start_time: + elapsed = current_time - self.episode_start_time + else: + elapsed = 0 + + # 记录时间戳 + self.timestamps.append(elapsed) + self.real_timestamps.append(datetime.now()) + + # 记录障碍物距离(从环境状态获取) + if len(current_state) > 0: + # 当前状态中的第一个元素是归一化的障碍物距离,需要反归一化 + norm_distance = current_state[0] + actual_distance = norm_distance * 300 + 300 # 反归一化 + self.obstacle_distances.append(actual_distance) + else: + self.obstacle_distances.append(0) + + # 记录动作 + self.actions.append(action) + + # 记录横向偏差和角度差 + if len(current_state) >= 4: + # 角度差(已归一化) + angle_error = current_state[2] # 这是phi + # 横向偏差(已归一化) + lateral_error = current_state[3] / 15 # 除以15反归一化 + + self.angle_errors.append(angle_error) + self.lateral_errors.append(lateral_error) + else: + self.angle_errors.append(0) + self.lateral_errors.append(0) + + # 记录速度(从车辆状态或环境状态获取) + if vehicle_state and 'speed_2d' in vehicle_state: + speed = vehicle_state['speed_2d'] + elif len(current_state) >= 2: + # 从状态计算速度(假设第二个状态是归一化速度) + norm_speed = (current_state[1] + current_state[0]) * 30 + 30 + speed = norm_speed / 3.6 # km/h转m/s + else: + speed = 0 + self.speeds.append(speed) + + # 计算并记录帧率 + if self.last_time: + frame_time = current_time - self.last_time + self.frame_times.append(frame_time) + if frame_time > 0: + fps = 1.0 / frame_time + self.fps_values.append(fps) + else: + self.fps_values.append(0) + + self.last_time = current_time + self.frame_count += 1 + self.rewards.append(reward) + + def end_episode(self): + """结束一个episode的数据收集""" + self.end_time = time.time() + episode_duration = self.end_time - self.episode_start_time if self.episode_start_time else 0 + + print(f"📈 Episode数据收集完成,总时长: {episode_duration:.2f}秒,记录步数: {len(self.timestamps)}") + + # 保存数据到CSV + self.save_to_csv() + + # 生成所有图表 + self.generate_all_plots() + + return episode_duration + + def save_to_csv(self): + """保存数据到CSV文件""" + if len(self.timestamps) == 0: + print("⚠️ 没有数据可保存") + return + + # 创建DataFrame + data = { + '时间戳': self.timestamps, + '实际时间': self.real_timestamps, + '障碍物距离': self.obstacle_distances, + '动作编号': self.actions, + '动作名称': [self.action_names_cn[a] if a < len(self.action_names_cn) else '未知' for a in self.actions], + '横向偏差': self.lateral_errors, + '角度差': self.angle_errors, + '速度_mps': self.speeds, + '速度_kmh': [s * 3.6 for s in self.speeds], + '帧率': self.fps_values, + '奖励值': self.rewards + } + + df = pd.DataFrame(data) + + # 生成文件名 + timestamp_str = datetime.now().strftime("%Y%m%d_%H%M%S") + filename = f"episode_data_{timestamp_str}.csv" + filepath = os.path.join(self.save_dir, filename) + + # 保存CSV + df.to_csv(filepath, index=False, encoding='utf-8-sig') + print(f"💾 数据已保存到: {filepath}") + + # 计算并打印统计信息 + self.print_statistics(df) + + def print_statistics(self, df): + """打印统计信息""" + print("\n" + "="*60) + print("数据统计") + print("="*60) + + if len(df) > 0: + print(f"总步数: {len(df)}") + print(f"平均障碍物距离: {df['障碍物距离'].mean():.2f}米") + print(f"最小障碍物距离: {df['障碍物距离'].min():.2f}米") + print(f"最大障碍物距离: {df['障碍物距离'].max():.2f}米") + print(f"平均横向偏差: {df['横向偏差'].abs().mean():.2f}米") + print(f"平均角度差: {df['角度差'].abs().mean():.2f}度") + print(f"平均速度: {df['速度_mps'].mean():.2f}米/秒 ({df['速度_kmh'].mean():.1f}公里/小时)") + print(f"平均帧率: {df['帧率'].mean():.1f} FPS") + + # 动作分布 + action_counts = df['动作名称'].value_counts() + print("\n动作分布:") + for action, count in action_counts.items(): + percentage = (count / len(df)) * 100 + print(f" {action}: {count}次 ({percentage:.1f}%)") + + def generate_all_plots(self): + """生成所有图表""" + if len(self.timestamps) == 0: + print("⚠️ 没有数据可绘制图表") + return + + print("📊 生成图表...") + + # 1. 障碍物距离与动作图 + self.plot_obstacle_distance_and_actions() + + # 2. 横向偏差和角度差折线图 + self.plot_lateral_and_angle_errors() + + # 3. 速度和帧率折线图 + self.plot_speed_and_fps() + + # 4. 综合图表 + self.plot_comprehensive_chart() + + print("✅ 所有图表已生成") + + def plot_obstacle_distance_and_actions(self): + """绘制障碍物距离与动作图""" + if len(self.timestamps) == 0: + return + + fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10)) + + # 如果中文字体不可用,使用简单的ASCII文本 + if not self.chinese_font_available: + print("⚠️ 中文字体不可用,图表可能无法正确显示中文") + + # 第一个子图:障碍物距离 + ax1.plot(self.timestamps, self.obstacle_distances, 'b-', linewidth=2, label='障碍物距离') + ax1.fill_between(self.timestamps, 0, self.obstacle_distances, alpha=0.2) + ax1.set_xlabel('运行时间 (秒)', fontsize=12) + ax1.set_ylabel('距离 (米)', fontsize=12) + ax1.set_title('障碍物距离随时间变化', fontsize=14, fontweight='bold') + ax1.grid(True, alpha=0.3) + ax1.legend(fontsize=10) + + # 添加安全距离线 + ax1.axhline(y=10, color='r', linestyle='--', alpha=0.5, label='安全距离(10米)') + ax1.axhline(y=5, color='orange', linestyle='--', alpha=0.5, label='警告距离(5米)') + + # 第二个子图:动作 + action_names = [self.action_names_cn[a] if a < len(self.action_names_cn) else '未知' for a in self.actions] + + # 使用散点图显示动作 + unique_actions = sorted(set(self.actions)) + for action in unique_actions: + indices = [i for i, a in enumerate(self.actions) if a == action] + times = [self.timestamps[i] for i in indices] + ax2.scatter(times, [action] * len(times), s=50, + label=self.action_names_cn[action] if action < len(self.action_names_cn) else '未知', + alpha=0.7) + + ax2.set_xlabel('运行时间 (秒)', fontsize=12) + ax2.set_ylabel('动作', fontsize=12) + ax2.set_title('执行动作分布', fontsize=14, fontweight='bold') + ax2.set_yticks(range(len(self.action_names_cn))) + ax2.set_yticklabels(self.action_names_cn) + ax2.grid(True, alpha=0.3) + ax2.legend(loc='upper right', fontsize=10) + + plt.tight_layout() + + # 保存图表 + timestamp_str = datetime.now().strftime("%Y%m%d_%H%M%S") + filename = f"障碍物距离与动作_{timestamp_str}.png" + filepath = os.path.join(self.plot_dir, filename) + plt.savefig(filepath, dpi=150, bbox_inches='tight') + plt.close() + + print(f" ✅ 障碍物距离与动作图已保存: {filename}") + + def plot_lateral_and_angle_errors(self): + """绘制横向偏差和角度差折线图""" + if len(self.timestamps) == 0: + return + + fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10)) + + # 第一个子图:横向偏差 + ax1.plot(self.timestamps, self.lateral_errors, 'g-', linewidth=2, label='横向偏差') + ax1.fill_between(self.timestamps, 0, self.lateral_errors, alpha=0.2, color='green') + ax1.set_xlabel('运行时间 (秒)', fontsize=12) + ax1.set_ylabel('横向偏差 (米)', fontsize=12) + ax1.set_title('横向偏差随时间变化', fontsize=14, fontweight='bold') + ax1.grid(True, alpha=0.3) + ax1.legend(fontsize=10) + + # 添加参考线 + ax1.axhline(y=0, color='k', linestyle='-', alpha=0.3) + ax1.axhline(y=1, color='orange', linestyle='--', alpha=0.5, label='允许偏差(1米)') + ax1.axhline(y=-1, color='orange', linestyle='--', alpha=0.5) + ax1.axhline(y=2, color='r', linestyle='--', alpha=0.5, label='最大偏差(2米)') + ax1.axhline(y=-2, color='r', linestyle='--', alpha=0.5) + + # 第二个子图:角度差 + ax2.plot(self.timestamps, self.angle_errors, 'r-', linewidth=2, label='角度差') + ax2.fill_between(self.timestamps, 0, self.angle_errors, alpha=0.2, color='red') + ax2.set_xlabel('运行时间 (秒)', fontsize=12) + ax2.set_ylabel('角度差 (度)', fontsize=12) + ax2.set_title('角度差随时间变化', fontsize=14, fontweight='bold') + ax2.grid(True, alpha=0.3) + ax2.legend(fontsize=10) + + # 添加参考线 + ax2.axhline(y=0, color='k', linestyle='-', alpha=0.3) + ax2.axhline(y=30, color='orange', linestyle='--', alpha=0.5, label='允许偏差(30°)') + ax2.axhline(y=-30, color='orange', linestyle='--', alpha=0.5) + ax2.axhline(y=100, color='r', linestyle='--', alpha=0.5, label='最大偏差(100°)') + ax2.axhline(y=-100, color='r', linestyle='--', alpha=0.5) + + plt.tight_layout() + + # 保存图表 + timestamp_str = datetime.now().strftime("%Y%m%d_%H%M%S") + filename = f"横向偏差与角度差_{timestamp_str}.png" + filepath = os.path.join(self.plot_dir, filename) + plt.savefig(filepath, dpi=150, bbox_inches='tight') + plt.close() + + print(f" ✅ 横向偏差和角度差图已保存: {filename}") + + def plot_speed_and_fps(self): + """绘制速度和帧率折线图""" + if len(self.timestamps) == 0: + return + + fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10)) + + # 第一个子图:速度 + speeds_kmh = [s * 3.6 for s in self.speeds] # 转换为km/h + + ax1.plot(self.timestamps, speeds_kmh, 'purple', linewidth=2, label='速度') + ax1.fill_between(self.timestamps, 0, speeds_kmh, alpha=0.2, color='purple') + ax1.set_xlabel('运行时间 (秒)', fontsize=12) + ax1.set_ylabel('速度 (公里/小时)', fontsize=12) + ax1.set_title('速度随时间变化', fontsize=14, fontweight='bold') + ax1.grid(True, alpha=0.3) + ax1.legend(fontsize=10) + + # 第二个子图:帧率 + ax2.plot(self.timestamps, self.fps_values, 'orange', linewidth=2, label='帧率') + ax2.fill_between(self.timestamps, 0, self.fps_values, alpha=0.2, color='orange') + ax2.set_xlabel('运行时间 (秒)', fontsize=12) + ax2.set_ylabel('帧率 (FPS)', fontsize=12) + ax2.set_title('帧率随时间变化', fontsize=14, fontweight='bold') + ax2.grid(True, alpha=0.3) + ax2.legend(fontsize=10) + + # 添加平均帧率线 + if len(self.fps_values) > 0: + avg_fps = np.mean(self.fps_values) + ax2.axhline(y=avg_fps, color='r', linestyle='--', alpha=0.7, + label=f'平均帧率: {avg_fps:.1f} FPS') + + plt.tight_layout() + + # 保存图表 + timestamp_str = datetime.now().strftime("%Y%m%d_%H%M%S") + filename = f"速度与帧率_{timestamp_str}.png" + filepath = os.path.join(self.plot_dir, filename) + plt.savefig(filepath, dpi=150, bbox_inches='tight') + plt.close() + + print(f" ✅ 速度和帧率图已保存: {filename}") + + def plot_comprehensive_chart(self): + """绘制综合图表""" + if len(self.timestamps) == 0: + return + + fig = plt.figure(figsize=(16, 12)) + + # 1. 障碍物距离 + ax1 = plt.subplot(3, 2, 1) + ax1.plot(self.timestamps, self.obstacle_distances, 'b-', linewidth=1.5) + ax1.set_xlabel('时间 (秒)', fontsize=10) + ax1.set_ylabel('距离 (米)', fontsize=10) + ax1.set_title('障碍物距离', fontsize=12, fontweight='bold') + ax1.grid(True, alpha=0.3) + + # 2. 横向偏差 + ax2 = plt.subplot(3, 2, 2) + ax2.plot(self.timestamps, self.lateral_errors, 'g-', linewidth=1.5) + ax2.set_xlabel('时间 (秒)', fontsize=10) + ax2.set_ylabel('横向偏差 (米)', fontsize=10) + ax2.set_title('横向偏差', fontsize=12, fontweight='bold') + ax2.grid(True, alpha=0.3) + + # 3. 角度差 + ax3 = plt.subplot(3, 2, 3) + ax3.plot(self.timestamps, self.angle_errors, 'r-', linewidth=1.5) + ax3.set_xlabel('时间 (秒)', fontsize=10) + ax3.set_ylabel('角度差 (度)', fontsize=10) + ax3.set_title('角度差', fontsize=12, fontweight='bold') + ax3.grid(True, alpha=0.3) + + # 4. 速度 + ax4 = plt.subplot(3, 2, 4) + speeds_kmh = [s * 3.6 for s in self.speeds] + ax4.plot(self.timestamps, speeds_kmh, 'purple', linewidth=1.5) + ax4.set_xlabel('时间 (秒)', fontsize=10) + ax4.set_ylabel('速度 (公里/小时)', fontsize=10) + ax4.set_title('速度', fontsize=12, fontweight='bold') + ax4.grid(True, alpha=0.3) + + # 5. 帧率 + ax5 = plt.subplot(3, 2, 5) + ax5.plot(self.timestamps, self.fps_values, 'orange', linewidth=1.5) + ax5.set_xlabel('时间 (秒)', fontsize=10) + ax5.set_ylabel('帧率 (FPS)', fontsize=10) + ax5.set_title('帧率', fontsize=12, fontweight='bold') + ax5.grid(True, alpha=0.3) + + # 6. 奖励值 + ax6 = plt.subplot(3, 2, 6) + ax6.plot(self.timestamps, self.rewards, 'brown', linewidth=1.5) + ax6.set_xlabel('时间 (秒)', fontsize=10) + ax6.set_ylabel('奖励值', fontsize=10) + ax6.set_title('奖励值', fontsize=12, fontweight='bold') + ax6.grid(True, alpha=0.3) + + plt.suptitle('自动驾驶性能综合图表', fontsize=16, fontweight='bold') + plt.tight_layout() + + # 保存图表 + timestamp_str = datetime.now().strftime("%Y%m%d_%H%M%S") + filename = f"综合图表_{timestamp_str}.png" + filepath = os.path.join(self.plot_dir, filename) + plt.savefig(filepath, dpi=150, bbox_inches='tight') + plt.close() + + print(f" ✅ 综合图表已保存: {filename}") + + def get_summary(self): + """获取数据摘要""" + summary = { + '总步数': len(self.timestamps), + '总时长': self.timestamps[-1] if self.timestamps else 0, + '平均障碍物距离': np.mean(self.obstacle_distances) if self.obstacle_distances else 0, + '最小障碍物距离': np.min(self.obstacle_distances) if self.obstacle_distances else 0, + '平均横向偏差': np.mean(np.abs(self.lateral_errors)) if self.lateral_errors else 0, + '平均角度差': np.mean(np.abs(self.angle_errors)) if self.angle_errors else 0, + '平均速度': np.mean(self.speeds) if self.speeds else 0, + '平均帧率': np.mean(self.fps_values) if self.fps_values else 0, + '总奖励值': np.sum(self.rewards) if self.rewards else 0 + } + + return summary + + def generate_performance_report(self, episode_num, episode_duration): + """生成性能报告""" + if len(self.timestamps) == 0: + return + + summary = self.get_summary() + + report = f""" + ================================================================ + Episode {episode_num} 性能报告 + ================================================================ + + 总体指标: + - 总步数: {summary['总步数']} + - 总时长: {summary['总时长']:.2f} 秒 + - 步速: {summary['总步数']/summary['总时长']:.2f} 步/秒 + + 安全指标: + - 平均障碍物距离: {summary['平均障碍物距离']:.2f} 米 + - 最小障碍物距离: {summary['最小障碍物距离']:.2f} 米 + - 平均横向偏差: {summary['平均横向偏差']:.2f} 米 + - 平均角度差: {summary['平均角度差']:.2f} 度 + + 性能指标: + - 平均速度: {summary['平均速度']*3.6:.1f} 公里/小时 + - 平均帧率: {summary['平均帧率']:.1f} FPS + - 总奖励值: {summary['总奖励值']:.1f} + + 数据保存: + - CSV文件: data_logs/episode_data_*.csv + - 图表: data_logs/plots/*.png + ================================================================ + """ + + print(report) + + # 保存报告到文件 + timestamp_str = datetime.now().strftime("%Y%m%d_%H%M%S") + report_file = os.path.join(self.save_dir, f"episode_{episode_num}_报告_{timestamp_str}.txt") + + with open(report_file, 'w', encoding='utf-8') as f: + f.write(report) + + print(f"📋 性能报告已保存到: {report_file}") \ No newline at end of file diff --git a/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/main.py b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/main.py index 5dab272650..89fff0b852 100644 --- a/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/main.py +++ b/src/Autonomous_vehicle_navigation_using_deep_learning_master/main/main.py @@ -1,8 +1,7 @@ #!/usr/bin/env python3 """ -主程序 - 简化版 +主程序 """ - import os os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' import time @@ -17,17 +16,46 @@ 'win-amd64' if os.name == 'nt' else 'linux-x86_64'))[0]) except IndexError: pass - import carla import numpy as np from collections import deque - +from data_collector import DataCollector from car_env import CarEnv from route_visualizer import RouteVisualizer from vehicle_tracker import VehicleTracker from traffic_manager import TrafficManager import config as cfg +import matplotlib +matplotlib.use('Agg') +import matplotlib.font_manager as fm + +def setup_chinese_font(): + """尝试设置中文字体""" + # 检查常见字体 + font_paths = [ + "C:/Windows/Fonts/simhei.ttf", + "/usr/share/fonts/truetype/wqy/wqy-microhei.ttc", + "/System/Library/Fonts/STHeiti Medium.ttc", + ] + + for font_path in font_paths: + if os.path.exists(font_path): + try: + fm.fontManager.addfont(font_path) + font_name = fm.FontProperties(fname=font_path).get_name() + matplotlib.rcParams['font.sans-serif'] = [font_name] + matplotlib.rcParams['axes.unicode_minus'] = False + print(f"✅ 使用中文字体: {font_name}") + return True + except: + continue + + print("⚠️ 未找到中文字体,图表可能无法显示中文") + return False + +setup_chinese_font() + def setup_environment(): """设置环境""" print("启动CARLA自动驾驶系统") @@ -113,6 +141,9 @@ def run_episode(env, traffic_mgr, visualizer, tracker, episode_num): """运行一个episode""" print(f"\nEpisode {episode_num}") + data_collector = DataCollector() + data_collector.start_episode() + # 加载模型 braking_model, driving_model = load_models() if not braking_model or not driving_model: @@ -175,6 +206,8 @@ def run_episode(env, traffic_mgr, visualizer, tracker, episode_num): print(f"执行动作失败: {e}") done = True + data_collector.record_step(env, action, current_state, reward, vehicle_state) + # 计算FPS frame_time = time.time() - step_start fps_counter.append(frame_time) @@ -186,6 +219,12 @@ def run_episode(env, traffic_mgr, visualizer, tracker, episode_num): if done: print(f"Episode {episode_num} 完成,步数: {step_count}") + episode_duration = data_collector.end_episode() + data_collector.generate_performance_report(episode_num, episode_duration) + # 打印项目完成时长 + print(f"\n⏱️ Episode {episode_num} 总时长: {episode_duration:.2f}秒") + print(f" 平均步速: {step_count/episode_duration:.2f} 步/秒") + print(f" 平均帧率: {data_collector.get_summary()['平均帧率']:.1f} FPS") break # 清理跟踪器资源