diff --git a/.idea/.gitignore b/.idea/.gitignore
new file mode 100644
index 0000000000..10b731c518
--- /dev/null
+++ b/.idea/.gitignore
@@ -0,0 +1,5 @@
+# 默认忽略的文件
+/shelf/
+/workspace.xml
+# 基于编辑器的 HTTP 客户端请求
+/httpRequests/
diff --git a/.idea/inspectionProfiles/profiles_settings.xml b/.idea/inspectionProfiles/profiles_settings.xml
new file mode 100644
index 0000000000..105ce2da2d
--- /dev/null
+++ b/.idea/inspectionProfiles/profiles_settings.xml
@@ -0,0 +1,6 @@
+
+
+
+
+
+
\ No newline at end of file
diff --git a/.idea/misc.xml b/.idea/misc.xml
new file mode 100644
index 0000000000..77773ffb78
--- /dev/null
+++ b/.idea/misc.xml
@@ -0,0 +1,7 @@
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/.idea/modules.xml b/.idea/modules.xml
new file mode 100644
index 0000000000..0e01bed9d6
--- /dev/null
+++ b/.idea/modules.xml
@@ -0,0 +1,8 @@
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/.idea/nn.iml b/.idea/nn.iml
new file mode 100644
index 0000000000..28462e1238
--- /dev/null
+++ b/.idea/nn.iml
@@ -0,0 +1,18 @@
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/.idea/vcs.xml b/.idea/vcs.xml
new file mode 100644
index 0000000000..35eb1ddfbb
--- /dev/null
+++ b/.idea/vcs.xml
@@ -0,0 +1,6 @@
+
+
+
+
+
+
\ No newline at end of file
diff --git a/src/Safe_navigation/README.md b/src/Safe_navigation/README.md
new file mode 100644
index 0000000000..e1ca3e8b66
--- /dev/null
+++ b/src/Safe_navigation/README.md
@@ -0,0 +1,2 @@
+首次提交选题为无人车安全导航
+
diff --git a/src/Safe_navigation/__pycache__/car.cpython-312-pytest-9.0.2.pyc b/src/Safe_navigation/__pycache__/car.cpython-312-pytest-9.0.2.pyc
new file mode 100644
index 0000000000..639510e718
Binary files /dev/null and b/src/Safe_navigation/__pycache__/car.cpython-312-pytest-9.0.2.pyc differ
diff --git a/src/Safe_navigation/__pycache__/connect_airsim.cpython-312-pytest-9.0.2.pyc b/src/Safe_navigation/__pycache__/connect_airsim.cpython-312-pytest-9.0.2.pyc
new file mode 100644
index 0000000000..448b5fae70
Binary files /dev/null and b/src/Safe_navigation/__pycache__/connect_airsim.cpython-312-pytest-9.0.2.pyc differ
diff --git a/src/Safe_navigation/car.py b/src/Safe_navigation/car.py
new file mode 100644
index 0000000000..5c8e3e4e70
--- /dev/null
+++ b/src/Safe_navigation/car.py
@@ -0,0 +1,846 @@
+"""
+基于深度强化学习(DQN)的AirSim自动驾驶安全导航训练系统
+兼容旧版AirSim服务器(版本1)的修复版
+"""
+
+import os
+import sys
+import time
+import random
+import argparse
+from collections import deque
+from datetime import datetime
+
+import numpy as np
+import cv2
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import torch.nn.functional as F
+from torch.utils.tensorboard import SummaryWriter
+
+# AirSim导入
+try:
+ import airsim
+ print(f"✓ AirSim模块导入成功,版本: {airsim.__version__}")
+except ImportError as e:
+ print(f"AirSim模块导入失败: {e}")
+ sys.exit(1)
+
+
+# ==================== 配置类 ====================
+class TrainingConfig:
+ """训练配置参数"""
+
+ def __init__(self):
+ # 网络参数
+ self.image_size = (84, 84) # 输入图像尺寸
+ self.state_dim = 15 # 简化状态维度
+ self.action_dim = 5 # 简化动作空间
+
+ # 训练参数
+ self.total_episodes = 50 # 测试阶段用少量回合
+ self.max_steps = 100
+ self.batch_size = 16
+ self.learning_rate = 1e-3
+ self.gamma = 0.99
+ self.tau = 1e-3
+ self.update_every = 4
+
+ # 经验回放
+ self.buffer_size = 2000
+ self.pretrain_length = 100
+
+ # 探索策略
+ self.epsilon_start = 1.0
+ self.epsilon_end = 0.1
+ self.epsilon_decay = 0.99
+
+ # 安全参数
+ self.collision_penalty = -5.0
+ self.max_speed = 10.0
+
+ # 路径参数
+ self.model_save_path = "./models"
+ self.log_path = "./logs"
+ self.save_interval = 10
+
+ # AirSim参数
+ self.ip_address = "127.0.0.1"
+
+
+# ==================== 神经网络架构 ====================
+class SimpleDQN(nn.Module):
+ """简化版DQN网络"""
+
+ def __init__(self, state_dim, action_dim, image_channels=3):
+ super(SimpleDQN, self).__init__()
+
+ # 视觉编码器
+ self.visual_encoder = nn.Sequential(
+ nn.Conv2d(image_channels, 8, kernel_size=4, stride=2),
+ nn.ReLU(),
+ nn.Conv2d(8, 16, kernel_size=3, stride=1),
+ nn.ReLU(),
+ nn.Flatten()
+ )
+
+ # 计算卷积层输出维度
+ with torch.no_grad():
+ sample = torch.zeros(1, image_channels, 84, 84)
+ conv_out = self.visual_encoder(sample)
+ self.visual_feature_dim = conv_out.shape[1]
+
+ # 状态处理器
+ self.state_processor = nn.Sequential(
+ nn.Linear(state_dim, 32),
+ nn.ReLU(),
+ )
+
+ # 特征融合层
+ fusion_input_dim = self.visual_feature_dim + 32
+ self.fusion_layer = nn.Sequential(
+ nn.Linear(fusion_input_dim, 64),
+ nn.ReLU(),
+ )
+
+ # 动作价值头
+ self.value_stream = nn.Sequential(
+ nn.Linear(64, 32),
+ nn.ReLU(),
+ nn.Linear(32, action_dim)
+ )
+
+ def forward(self, image, state):
+ # 处理视觉输入
+ visual_features = self.visual_encoder(image)
+
+ # 处理状态输入
+ state_features = self.state_processor(state)
+
+ # 特征融合
+ combined = torch.cat([visual_features, state_features], dim=1)
+ fused_features = self.fusion_layer(combined)
+
+ # 输出动作价值
+ q_values = self.value_stream(fused_features)
+
+ return q_values
+
+
+# ==================== 经验回放缓冲区 ====================
+class ReplayBuffer:
+ """简化版经验回放缓冲区"""
+
+ def __init__(self, capacity):
+ self.buffer = deque(maxlen=capacity)
+
+ def push(self, experience):
+ self.buffer.append(experience)
+
+ def sample(self, batch_size):
+ if len(self.buffer) < batch_size:
+ return None
+
+ indices = np.random.choice(len(self.buffer), batch_size, replace=False)
+ return [self.buffer[idx] for idx in indices]
+
+ def __len__(self):
+ return len(self.buffer)
+
+
+# ==================== DQN智能体 ====================
+class DQNAgent:
+ """DQN智能体"""
+
+ def __init__(self, config, device='cpu'):
+ self.config = config
+ self.device = torch.device(device)
+
+ # 初始化网络
+ self.policy_net = SimpleDQN(config.state_dim, config.action_dim).to(self.device)
+ self.target_net = SimpleDQN(config.state_dim, config.action_dim).to(self.device)
+ self.target_net.load_state_dict(self.policy_net.state_dict())
+ self.target_net.eval()
+
+ # 优化器
+ self.optimizer = optim.Adam(self.policy_net.parameters(), lr=config.learning_rate)
+
+ # 经验回放缓冲区
+ self.memory = ReplayBuffer(config.buffer_size)
+
+ # 训练参数
+ self.epsilon = config.epsilon_start
+ self.steps_done = 0
+
+ print(f"初始化DQN智能体,设备: {self.device}")
+
+ def select_action(self, state_image, state_vector, eval_mode=False):
+ """选择动作"""
+ if not eval_mode and random.random() < self.epsilon:
+ return random.randrange(self.config.action_dim)
+
+ with torch.no_grad():
+ image_tensor = torch.FloatTensor(state_image).unsqueeze(0).to(self.device)
+ vector_tensor = torch.FloatTensor(state_vector).unsqueeze(0).to(self.device)
+
+ q_values = self.policy_net(image_tensor, vector_tensor)
+ return q_values.argmax(1).item()
+
+ def train_step(self):
+ """训练步骤"""
+ if len(self.memory) < self.config.pretrain_length:
+ return 0
+
+ batch = self.memory.sample(self.config.batch_size)
+ if batch is None:
+ return 0
+
+ # 解构批数据
+ states_img, states_vec, actions, rewards, next_states_img, next_states_vec, dones = zip(*batch)
+
+ # 转换为张量
+ states_img = torch.FloatTensor(np.array(states_img)).to(self.device)
+ states_vec = torch.FloatTensor(np.array(states_vec)).to(self.device)
+ actions = torch.LongTensor(actions).unsqueeze(1).to(self.device)
+ rewards = torch.FloatTensor(rewards).unsqueeze(1).to(self.device)
+ next_states_img = torch.FloatTensor(np.array(next_states_img)).to(self.device)
+ next_states_vec = torch.FloatTensor(np.array(next_states_vec)).to(self.device)
+ dones = torch.FloatTensor(dones).unsqueeze(1).to(self.device)
+
+ # 计算当前Q值
+ current_q = self.policy_net(states_img, states_vec)
+ current_q = current_q.gather(1, actions)
+
+ # 计算目标Q值
+ with torch.no_grad():
+ next_q = self.target_net(next_states_img, next_states_vec)
+ next_q_max = next_q.max(1)[0].unsqueeze(1)
+ target_q = rewards + (1 - dones) * self.config.gamma * next_q_max
+
+ # 计算损失
+ loss = F.smooth_l1_loss(current_q, target_q)
+
+ # 反向传播
+ self.optimizer.zero_grad()
+ loss.backward()
+ torch.nn.utils.clip_grad_norm_(self.policy_net.parameters(), 1.0)
+ self.optimizer.step()
+
+ # 更新探索率
+ self.epsilon = max(self.config.epsilon_end, self.epsilon * self.config.epsilon_decay)
+
+ return loss.item()
+
+ def update_target_network(self):
+ """更新目标网络"""
+ self.target_net.load_state_dict(self.policy_net.state_dict())
+
+ def save_model(self, episode, path=None):
+ """保存模型"""
+ if path is None:
+ path = self.config.model_save_path
+
+ os.makedirs(path, exist_ok=True)
+ model_path = os.path.join(path, f'airsim_dqn_episode_{episode}.pth')
+
+ torch.save({
+ 'episode': episode,
+ 'policy_state_dict': self.policy_net.state_dict(),
+ 'optimizer_state_dict': self.optimizer.state_dict(),
+ 'epsilon': self.epsilon,
+ }, model_path)
+
+ print(f"模型已保存: {model_path}")
+
+ def load_model(self, model_path):
+ """加载模型"""
+ checkpoint = torch.load(model_path, map_location=self.device)
+
+ self.policy_net.load_state_dict(checkpoint['policy_state_dict'])
+ self.optimizer.load_state_dict(checkpoint['optimizer_state_dict'])
+ self.epsilon = checkpoint['epsilon']
+
+ print(f"模型已加载: {model_path}")
+
+
+# ==================== AirSim环境封装(兼容旧版) ====================
+class AirSimSafetyEnv:
+ """兼容旧版AirSim服务器的环境封装"""
+
+ def __init__(self, config):
+ self.config = config
+ self.client = None
+ self.step_count = 0
+ self.total_reward = 0
+ self.collisions = 0
+
+ print("初始化AirSim环境...")
+
+ def connect(self):
+ """连接到AirSim服务器(兼容旧版)"""
+ try:
+ print(f"尝试连接AirSim服务器 {self.config.ip_address}...")
+
+ # 创建客户端
+ self.client = airsim.CarClient()
+ self.client.confirmConnection()
+ print("✓ 连接成功!")
+
+ # 尝试启用API控制(兼容旧版)
+ try:
+ self.client.enableApiControl(True)
+ print("✓ API控制已启用")
+ except Exception as api_error:
+ print(f"⚠️ API控制启用失败(可能是旧版): {api_error}")
+ print("尝试继续运行...")
+
+ # 重置车辆
+ self.client.reset()
+ print("✓ 车辆已重置")
+
+ # 等待稳定
+ time.sleep(1.0)
+
+ return True
+
+ except Exception as e:
+ print(f"❌ 连接失败: {e}")
+ print("\n请确保:")
+ print("1. AirSim仿真环境正在运行(如AirSimNH.exe)")
+ print("2. 已选择汽车模式")
+ print("3. 服务器版本兼容(使用AirSim 1.2.6客户端)")
+ return False
+
+ def get_camera_image(self):
+ """获取摄像头图像"""
+ try:
+ # 获取场景图像
+ responses = self.client.simGetImages([
+ airsim.ImageRequest("0", airsim.ImageType.Scene, False, False)
+ ])
+
+ if responses and len(responses) > 0:
+ img1d = np.frombuffer(responses[0].image_data_uint8, dtype=np.uint8)
+
+ if img1d.size > 0:
+ img_rgb = img1d.reshape(responses[0].height, responses[0].width, 3)
+ resized = cv2.resize(img_rgb, self.config.image_size)
+ normalized = resized.astype(np.float32) / 255.0
+ normalized = np.transpose(normalized, (2, 0, 1))
+
+ return normalized
+
+ except Exception as e:
+ print(f"获取摄像头图像失败: {e}")
+
+ # 返回空白图像
+ return np.zeros((3, *self.config.image_size), dtype=np.float32)
+
+ def get_vehicle_state(self):
+ """获取车辆状态"""
+ try:
+ # 获取车辆状态
+ car_state = self.client.getCarState()
+
+ # 获取碰撞信息
+ collision_info = self.client.simGetCollisionInfo()
+
+ state_info = {
+ 'speed': car_state.speed,
+ 'velocity': [
+ car_state.kinematics_estimated.linear_velocity.x_val,
+ car_state.kinematics_estimated.linear_velocity.y_val,
+ car_state.kinematics_estimated.linear_velocity.z_val
+ ],
+ 'position': [
+ car_state.kinematics_estimated.position.x_val,
+ car_state.kinematics_estimated.position.y_val,
+ car_state.kinematics_estimated.position.z_val
+ ],
+ 'collision': collision_info.has_collided,
+ 'collision_count': collision_info.collision_count,
+ }
+
+ return state_info
+
+ except Exception as e:
+ print(f"获取车辆状态失败: {e}")
+ return None
+
+ def create_state_vector(self, state_info):
+ """创建状态向量"""
+ if state_info is None:
+ return np.zeros(self.config.state_dim, dtype=np.float32)
+
+ state_vector = []
+
+ # 速度信息
+ state_vector.append(state_info['speed'] / self.config.max_speed)
+ state_vector.extend([v / 10.0 for v in state_info['velocity'][:2]])
+
+ # 位置信息
+ state_vector.extend([p / 100.0 for p in state_info['position'][:2]])
+
+ # 碰撞信息
+ state_vector.append(float(state_info['collision']))
+
+ # 补全到指定维度
+ while len(state_vector) < self.config.state_dim:
+ state_vector.append(np.random.uniform(-0.1, 0.1))
+
+ state_vector = state_vector[:self.config.state_dim]
+
+ return np.array(state_vector, dtype=np.float32)
+
+ def get_state(self):
+ """获取当前状态"""
+ try:
+ image_state = self.get_camera_image()
+ state_info = self.get_vehicle_state()
+ state_vector = self.create_state_vector(state_info)
+
+ safety_flags = {
+ 'collision': state_info['collision'] if state_info else False,
+ }
+
+ return image_state, state_vector, safety_flags
+
+ except Exception as e:
+ print(f"获取状态失败: {e}")
+ # 返回默认状态
+ image_state = np.zeros((3, *self.config.image_size), dtype=np.float32)
+ state_vector = np.zeros(self.config.state_dim, dtype=np.float32)
+ safety_flags = {'collision': False}
+ return image_state, state_vector, safety_flags
+
+ def apply_action(self, action_idx):
+ """应用动作到车辆"""
+ # 简化动作空间
+ steer_actions = [-0.3, -0.1, 0.0, 0.1, 0.3]
+ throttle_actions = [0.0, 0.2, 0.5, 0.8, 1.0]
+
+ # 确保动作索引在范围内
+ action_idx = min(action_idx, len(steer_actions) * len(throttle_actions) - 1)
+ steer_idx = action_idx % len(steer_actions)
+ throttle_idx = min(action_idx // len(steer_actions), len(throttle_actions) - 1)
+
+ # 创建控制命令
+ car_controls = airsim.CarControls()
+ car_controls.steering = steer_actions[steer_idx]
+ car_controls.throttle = throttle_actions[throttle_idx]
+ car_controls.brake = 0.0
+
+ # 应用控制
+ try:
+ self.client.setCarControls(car_controls)
+ return car_controls
+ except Exception as e:
+ print(f"应用控制命令失败: {e}")
+ return car_controls
+
+ def calculate_reward(self, current_state_info, safety_flags):
+ """计算奖励函数"""
+ if current_state_info is None:
+ return 0.0
+
+ reward = 0.0
+ speed = current_state_info['speed']
+
+ # 基础移动奖励
+ if speed > 0.1:
+ reward += 0.1
+
+ # 碰撞惩罚
+ if safety_flags['collision']:
+ reward += self.config.collision_penalty
+ self.collisions += 1
+ print(f"⚠️ 发生碰撞! 惩罚: {self.config.collision_penalty}")
+
+ # 生存奖励
+ reward += 0.01
+
+ return reward
+
+ def step(self, action_idx):
+ """执行一步环境交互"""
+ self.step_count += 1
+
+ # 获取当前状态
+ prev_image, prev_vector, prev_safety = self.get_state()
+
+ # 应用动作
+ control = self.apply_action(action_idx)
+
+ # 等待环境响应
+ time.sleep(0.1)
+
+ # 获取新状态
+ current_image, current_vector, current_safety = self.get_state()
+
+ # 获取状态信息用于计算奖励
+ current_state_info = self.get_vehicle_state()
+
+ # 计算奖励
+ reward = self.calculate_reward(current_state_info, current_safety)
+ self.total_reward += reward
+
+ # 检查是否终止
+ done = False
+ if current_safety['collision']:
+ done = True
+ print("💥 终止: 发生碰撞")
+ elif self.step_count >= self.config.max_steps:
+ done = True
+ print("⏱️ 终止: 达到最大步数")
+
+ return (current_image, current_vector, reward,
+ prev_image, prev_vector, done, current_safety)
+
+ def reset(self):
+ """重置环境"""
+ self.step_count = 0
+ self.total_reward = 0
+ self.collisions = 0
+
+ try:
+ self.client.reset()
+ time.sleep(1.0) # 等待重置完成
+ except Exception as e:
+ print(f"重置环境失败: {e}")
+
+ image_state, vector_state, safety_flags = self.get_state()
+
+ return image_state, vector_state, safety_flags
+
+ def close(self):
+ """关闭环境"""
+ try:
+ print("AirSim环境已关闭")
+ except:
+ pass
+
+
+# ==================== 训练函数 ====================
+def train_dqn_safety_navigation(resume_model=None):
+ """主训练函数"""
+ config = TrainingConfig()
+
+ # 创建TensorBoard记录器
+ log_dir = f"./logs/airsim_dqn_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
+ writer = SummaryWriter(log_dir)
+
+ # 初始化环境和智能体
+ env = AirSimSafetyEnv(config)
+ agent = DQNAgent(config, device='cpu')
+
+ # 恢复训练(如果指定)
+ if resume_model and os.path.exists(resume_model):
+ try:
+ agent.load_model(resume_model)
+ print(f"从模型恢复训练: {resume_model}")
+ except:
+ print("模型加载失败,从头开始训练")
+ else:
+ print("从头开始训练")
+
+ # 连接AirSim服务器
+ print("\n" + "=" * 60)
+ print("正在连接AirSim服务器...")
+
+ if not env.connect():
+ print("无法连接到AirSim服务器,退出训练")
+ return
+
+ print("=" * 60)
+
+ # 训练循环
+ print(f"\n开始深度强化学习安全导航训练")
+ print(f"总回合数: {config.total_episodes}")
+ print(f"最大步数: {config.max_steps}")
+ print(f"设备: {agent.device}")
+ print("=" * 60)
+
+ for episode in range(config.total_episodes):
+ # 重置环境
+ try:
+ image_state, vector_state, safety_flags = env.reset()
+ except Exception as e:
+ print(f"重置环境失败: {e}")
+ break
+
+ episode_reward = 0
+ episode_steps = 0
+ episode_losses = []
+ episode_start_time = time.time()
+
+ # 回合循环
+ done = False
+ while not done and episode_steps < config.max_steps:
+ try:
+ # 选择动作
+ action = agent.select_action(image_state, vector_state)
+
+ # 执行动作,获取新状态和奖励
+ (next_image, next_vector, reward,
+ prev_image, prev_vector, done, next_safety) = env.step(action)
+
+ # 存储经验
+ experience = (
+ prev_image, prev_vector, action, reward,
+ next_image, next_vector, done
+ )
+ agent.memory.push(experience)
+
+ # 训练智能体
+ if agent.steps_done % config.update_every == 0:
+ loss = agent.train_step()
+ if loss > 0:
+ episode_losses.append(loss)
+
+ # 更新状态
+ image_state, vector_state = next_image, next_vector
+ episode_reward += reward
+ episode_steps += 1
+ agent.steps_done += 1
+
+ # 简单进度显示
+ if episode_steps % 10 == 0:
+ print(f" 步数: {episode_steps}, 奖励: {episode_reward:.2f}, 探索率: {agent.epsilon:.3f}")
+
+ except Exception as e:
+ print(f"回合执行出错: {e}")
+ done = True
+
+ # 计算回合统计
+ episode_time = time.time() - episode_start_time
+ avg_loss = np.mean(episode_losses) if episode_losses else 0
+
+ # 记录训练数据
+ writer.add_scalar('Reward/Episode', episode_reward, episode)
+ writer.add_scalar('Loss/Episode', avg_loss, episode)
+ writer.add_scalar('Exploration/Epsilon', agent.epsilon, episode)
+ writer.add_scalar('Steps/Episode_Steps', episode_steps, episode)
+
+ # 打印回合总结
+ print(f"\n回合 {episode + 1}/{config.total_episodes}")
+ print(f" 总奖励: {episode_reward:.2f}")
+ print(f" 步数: {episode_steps}")
+ print(f" 时间: {episode_time:.1f}s")
+ print(f" 平均损失: {avg_loss:.4f}")
+ print(f" 碰撞次数: {env.collisions}")
+ print(f" 探索率: {agent.epsilon:.3f}")
+
+ # 保存模型
+ if (episode + 1) % config.save_interval == 0:
+ agent.save_model(episode + 1)
+
+ # 更新目标网络
+ if (episode + 1) % 5 == 0:
+ agent.update_target_network()
+
+ # 保存最终模型
+ agent.save_model(config.total_episodes)
+
+ # 关闭环境
+ env.close()
+ writer.close()
+
+ print("\n" + "=" * 60)
+ print("🎉 训练完成!")
+ print(f"模型已保存至: {config.model_save_path}")
+ print(f"训练日志: {log_dir}")
+ print("=" * 60)
+
+
+# ==================== 评估函数 ====================
+def evaluate_model(model_path, eval_episodes=3):
+ """评估训练好的模型"""
+ config = TrainingConfig()
+ env = AirSimSafetyEnv(config)
+ agent = DQNAgent(config, device='cpu')
+
+ # 加载模型
+ try:
+ agent.load_model(model_path)
+ agent.epsilon = 0.01 # 评估时探索率低
+ except Exception as e:
+ print(f"模型加载失败: {e}")
+ return
+
+ # 连接环境
+ if not env.connect():
+ print("无法连接到AirSim服务器")
+ return
+
+ results = []
+
+ for episode in range(eval_episodes):
+ # 重置环境
+ image_state, vector_state, _ = env.reset()
+
+ episode_reward = 0
+ episode_steps = 0
+
+ done = False
+ while not done and episode_steps < config.max_steps:
+ try:
+ # 选择动作(评估模式)
+ action = agent.select_action(image_state, vector_state, eval_mode=True)
+
+ # 执行动作
+ (next_image, next_vector, reward,
+ _, _, done, _) = env.step(action)
+
+ # 更新
+ image_state, vector_state = next_image, next_vector
+ episode_reward += reward
+ episode_steps += 1
+
+ # 简单显示
+ if episode_steps % 10 == 0:
+ print(f" 评估步数: {episode_steps}, 奖励: {episode_reward:.2f}")
+
+ except Exception as e:
+ print(f"评估出错: {e}")
+ done = True
+
+ results.append({
+ 'episode': episode + 1,
+ 'reward': episode_reward,
+ 'steps': episode_steps,
+ })
+
+ print(f"评估回合 {episode + 1}/{eval_episodes}: 奖励={episode_reward:.2f}")
+
+ # 计算平均指标
+ if results:
+ avg_reward = np.mean([r['reward'] for r in results])
+
+ print("\n" + "=" * 60)
+ print("评估结果总结:")
+ print(f"平均回合奖励: {avg_reward:.2f}")
+ print("=" * 60)
+
+ env.close()
+
+
+# ==================== 主程序入口 ====================
+if __name__ == "__main__":
+ import warnings
+ warnings.filterwarnings('ignore')
+
+ print("=" * 70)
+ print("基于深度强化学习的AirSim自动驾驶安全导航系统")
+ print("版本:兼容旧版AirSim服务器 (1.2.6)")
+ print("=" * 70)
+
+ # 创建必要的目录
+ os.makedirs('./models', exist_ok=True)
+ os.makedirs('./logs', exist_ok=True)
+
+ # 检查PyTorch
+ print(f"PyTorch版本: {torch.__version__}")
+ if torch.cuda.is_available():
+ print(f"检测到GPU: {torch.cuda.get_device_name(0)}")
+ else:
+ print("将使用CPU运行")
+
+ # 用户选择
+ choice = None
+ valid_choices = ['1', '2', '3']
+
+ while choice not in valid_choices:
+ print("\n请选择运行模式:")
+ print("1. 训练新模型 (推荐先测试连接)")
+ print("2. 恢复训练")
+ print("3. 评估模型")
+
+ user_input = input("\n请输入选择 (1-3): ").strip()
+
+ if user_input:
+ first_char = user_input[0]
+ if first_char in valid_choices:
+ choice = first_char
+ else:
+ print(f"错误: 输入 '{user_input}' 无效。请输入 1, 2 或 3。")
+ else:
+ print("错误: 输入不能为空。请输入 1, 2 或 3。")
+
+ # 根据选择执行
+ if choice == '1':
+ print("\n" + "=" * 60)
+ print("重要提示:")
+ print("1. 请确保AirSim仿真环境正在运行")
+ print("2. 已选择汽车模式 (Car Mode)")
+ print("3. 初始训练只有50回合,用于测试连接")
+ print("=" * 60)
+
+ confirm = input("\n确认AirSim环境已启动?(y/n): ").strip().lower()
+ if confirm == 'y':
+ train_dqn_safety_navigation()
+ else:
+ print("请先启动AirSim环境再运行程序")
+ elif choice == '2':
+ model_path = None
+ while not model_path:
+ model_path = input("请输入模型路径 (例如: ./models/airsim_dqn_episode_10.pth): ").strip()
+ if not model_path:
+ print("错误: 输入不能为空。")
+ continue
+
+ if os.path.exists(model_path):
+ print(f"从模型恢复训练: {model_path}")
+ train_dqn_safety_navigation(resume_model=model_path)
+ else:
+ print(f"错误: 模型文件不存在: {model_path}")
+ model_path = None
+ elif choice == '3':
+ model_path = None
+ while not model_path:
+ model_path = input("请输入要评估的模型路径: ").strip()
+ if not model_path:
+ print("错误: 输入不能为空。")
+ continue
+
+ if os.path.exists(model_path):
+ print(f"评估模型: {model_path}")
+ evaluate_model(model_path)
+ else:
+ print(f"错误: 模型文件不存在: {model_path}")
+ model_path = None
+
+
+ # 在代码开头添加测试函数
+ def test_airsim_connection():
+ """测试AirSim连接"""
+ try:
+ import airsim
+ client = airsim.CarClient()
+ client.confirmConnection()
+ print("✓ 成功连接到AirSim服务器!")
+
+ # 获取车辆状态
+ state = client.getCarState()
+ print(f"车辆速度: {state.speed}")
+
+ client.enableApiControl(True)
+ print("API控制已启用")
+
+ return True
+ except Exception as e:
+ print(f"❌ 连接失败: {e}")
+ print("\n请确保:")
+ print("1. AirSim仿真环境正在运行")
+ print("2. 已选择汽车模式")
+ print("3. AirSim服务器IP地址正确")
+ return False
+
+
+ # 在主程序中调用
+ if __name__ == "__main__":
+ print("测试AirSim连接...")
+ if test_airsim_connection():
+ print("\n连接测试通过!可以开始训练。")
+ else:
+ print("\n连接测试失败,请检查AirSim环境。")
\ No newline at end of file
diff --git a/src/Safe_navigation/check_version.py b/src/Safe_navigation/check_version.py
new file mode 100644
index 0000000000..d36e7ce000
--- /dev/null
+++ b/src/Safe_navigation/check_version.py
@@ -0,0 +1,30 @@
+import airsim
+import pkg_resources
+import inspect
+
+print("=== AirSim 环境诊断 ===")
+
+# 1. 检查安装的版本
+try:
+ version = pkg_resources.get_distribution("airsim").version
+ print(f"1. 已安装的 airsim 包版本: {version}")
+except Exception as e:
+ print(f"1. 无法获取版本: {e}")
+
+# 2. 查看 CarClient 的构造函数签名
+print("\n2. CarClient.__init__ 接受的参数:")
+try:
+ # 获取构造函数签名,跳过第一个参数'self'
+ sig = inspect.signature(airsim.CarClient.__init__)
+ params = list(sig.parameters.keys())[1:] # 移除'self'
+ print(" ", params)
+except Exception as e:
+ print(f" 获取失败: {e}")
+
+# 3. 尝试导入其他可能相关的模块
+print("\n3. 尝试直接导入常用客户端...")
+try:
+ from airsim import CarClient
+ print(" ✓ 成功从 airsim 导入 CarClient")
+except ImportError as e:
+ print(f" ✗ 导入失败: {e}")
\ No newline at end of file
diff --git a/src/Safe_navigation/connect_airsim.py b/src/Safe_navigation/connect_airsim.py
new file mode 100644
index 0000000000..f1561402f2
--- /dev/null
+++ b/src/Safe_navigation/connect_airsim.py
@@ -0,0 +1,51 @@
+import airsim
+import time
+
+
+def main():
+ print("=== 尝试连接 AirSimNH (使用 airsim 1.8.1) ===")
+ print("重要:请先确保虚幻引擎中的 AirSimNH 已点击播放(Play)!\n")
+
+ try:
+ # 1. 创建客户端(1.8.1版本的唯一正确方式)
+ client = airsim.CarClient()
+ print("✓ 客户端对象创建成功")
+
+ # 2. 确认连接(这会尝试与仿真器通信)
+ client.confirmConnection()
+ print("✓ 已连接到AirSim仿真服务器")
+
+ # 3. 启用控制
+ client.enableApiControl(True)
+ print("✓ API控制已启用")
+
+ # 4. 获取车辆状态,验证一切正常
+ car_state = client.getCarState()
+ print(f"✓ 车辆状态获取成功 - 速度: {car_state.speed} km/h")
+
+ # 5. 【可选】简单控制演示
+ print("\n>>> 连接成功!开始简单控制演示...")
+ controls = airsim.CarControls()
+ controls.throttle = 0.5
+ client.setCarControls(controls)
+ time.sleep(2)
+ controls.brake = 1.0
+ controls.throttle = 0.0
+ client.setCarControls(controls)
+ time.sleep(1)
+ print("演示结束。")
+ # 6. 释放控制
+ client.enableApiControl(False)
+ print("控制权已释放。")
+
+ except ConnectionRefusedError:
+ print("\n✗ 连接被拒绝。")
+ print(" 最可能的原因:虚幻引擎中的 AirSimNH 仿真没有启动。")
+ print(" 请打开虚幻引擎,加载AirSimNH项目,并点击顶部工具栏的蓝色【播放】(▶)按钮。")
+ except Exception as e:
+ print(f"\n✗ 连接过程中出错: {e}")
+ print(" 其他可能原因:防火墙阻止、端口占用或配置文件错误。")
+
+
+if __name__ == "__main__":
+ main()
\ No newline at end of file
diff --git a/src/Safe_navigation/logs/airsim_dqn_20251221_193518/events.out.tfevents.1766316918.YD.1800.0 b/src/Safe_navigation/logs/airsim_dqn_20251221_193518/events.out.tfevents.1766316918.YD.1800.0
new file mode 100644
index 0000000000..039f139958
Binary files /dev/null and b/src/Safe_navigation/logs/airsim_dqn_20251221_193518/events.out.tfevents.1766316918.YD.1800.0 differ
diff --git a/src/Safe_navigation/run_main.py b/src/Safe_navigation/run_main.py
new file mode 100644
index 0000000000..ba4c30b1ac
--- /dev/null
+++ b/src/Safe_navigation/run_main.py
@@ -0,0 +1,506 @@
+#!/usr/bin/env python3
+"""
+AirSimNH 无人车完整仿真控制脚本
+功能:连接仿真器、手动控制车辆、采集传感器数据、监控车辆状态
+"""
+
+import airsim
+import time
+import numpy as np
+import cv2
+import json
+import os
+from datetime import datetime
+import threading
+from collections import deque
+
+
+class AirSimNHCarSimulator:
+ """AirSim无人车仿真主类"""
+
+ def __init__(self, ip="127.0.0.1", port=41451, vehicle_name="PhysXCar"):
+ """
+ 初始化仿真器连接
+
+ 参数:
+ ip: AirSim服务器IP地址
+ port: AirSim服务器端口
+ vehicle_name: 车辆名称,需与settings.json中一致
+ """
+ self.ip = ip
+ self.port = port
+ self.vehicle_name = vehicle_name
+ self.client = None
+ self.is_connected = False
+ self.is_api_control_enabled = False
+ self.running = False
+ self.data_log = []
+ self.data_file = None
+
+ # 传感器数据缓存
+ self.sensor_data = {
+ "camera": deque(maxlen=100),
+ "imu": deque(maxlen=1000),
+ "gps": deque(maxlen=1000),
+ "lidar": deque(maxlen=100)
+ }
+
+ # 创建数据保存目录
+ timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+ self.data_dir = f"simulation_data_{timestamp}"
+ os.makedirs(self.data_dir, exist_ok=True)
+
+ print(f"数据保存目录: {self.data_dir}")
+
+ def connect(self):
+ """连接到AirSim仿真器"""
+ try:
+ print(f"正在连接到AirSim仿真器 {self.ip}:{self.port}...")
+
+ # 创建客户端连接
+ self.client = airsim.CarClient(ip=self.ip, port=self.port)
+ self.client.confirmConnection()
+
+ # 检查车辆是否存在
+ vehicles = self.client.listVehicles()
+ if self.vehicle_name not in vehicles:
+ print(f"警告: 车辆 '{self.vehicle_name}' 未找到,可用车辆: {vehicles}")
+ # 尝试使用找到的第一个车辆
+ if vehicles:
+ self.vehicle_name = vehicles[0]
+ print(f"使用车辆: {self.vehicle_name}")
+
+ self.is_connected = True
+ print("✓ 成功连接到AirSim仿真器!")
+ return True
+
+ except Exception as e:
+ print(f"✗ 连接失败: {e}")
+ print("请确保:")
+ print("1. AirSimNH环境正在运行 (在虚幻引擎中启动)")
+ print("2. settings.json配置正确")
+ print("3. 网络连接正常")
+ return False
+
+ def enable_api_control(self, enable=True):
+ """启用/禁用API控制"""
+ try:
+ self.client.enableApiControl(enable, vehicle_name=self.vehicle_name)
+ self.is_api_control_enabled = enable
+
+ if enable:
+ print("✓ API控制已启用")
+ # 重置控制到初始状态
+ controls = airsim.CarControls()
+ controls.throttle = 0
+ controls.steering = 0
+ controls.brake = 0
+ controls.handbrake = False
+ self.client.setCarControls(controls, vehicle_name=self.vehicle_name)
+ else:
+ print("✓ API控制已禁用")
+
+ return True
+
+ except Exception as e:
+ print(f"✗ API控制设置失败: {e}")
+ return False
+
+ def get_vehicle_state(self):
+ """获取完整的车辆状态信息"""
+ try:
+ state = self.client.getCarState(vehicle_name=self.vehicle_name)
+
+ # 获取车辆物理信息
+ kinematics = self.client.simGetVehiclePose(vehicle_name=self.vehicle_name)
+
+ state_info = {
+ "timestamp": time.time(),
+ "speed_kmh": state.speed,
+ "speed_ms": state.speed / 3.6,
+ "position": {
+ "x": kinematics.position.x_val,
+ "y": kinematics.position.y_val,
+ "z": kinematics.position.z_val
+ },
+ "orientation": {
+ "w": kinematics.orientation.w_val,
+ "x": kinematics.orientation.x_val,
+ "y": kinematics.orientation.y_val,
+ "z": kinematics.orientation.z_val
+ },
+ "gear": state.gear,
+ "rpm": state.rpm,
+ "max_rpm": state.maxrpm,
+ "handbrake": state.handbrake,
+ "collision": state.collision.has_collided,
+ "collision_count": state.collision.collision_count
+ }
+
+ return state_info
+
+ except Exception as e:
+ print(f"获取车辆状态失败: {e}")
+ return None
+
+ def capture_camera_images(self, camera_names=["front", "back", "left", "right"]):
+ """从多个摄像头捕获图像"""
+ images = {}
+
+ for cam_name in camera_names:
+ try:
+ # 请求RGB图像
+ responses = self.client.simGetImages([
+ airsim.ImageRequest(cam_name, airsim.ImageType.Scene, False, False)
+ ], vehicle_name=self.vehicle_name)
+
+ if responses and responses[0]:
+ img_response = responses[0]
+
+ # 将图像数据转换为numpy数组
+ img1d = np.frombuffer(img_response.image_data_uint8, dtype=np.uint8)
+ img_rgb = img1d.reshape(img_response.height, img_response.width, 3)
+
+ # 保存图像到文件
+ timestamp = datetime.now().strftime("%H%M%S_%f")[:-3]
+ filename = f"{self.data_dir}/{cam_name}_{timestamp}.png"
+ cv2.imwrite(filename, cv2.cvtColor(img_rgb, cv2.COLOR_RGB2BGR))
+
+ images[cam_name] = {
+ "filename": filename,
+ "shape": img_rgb.shape,
+ "timestamp": time.time()
+ }
+
+ # 缓存数据
+ self.sensor_data["camera"].append({
+ "camera": cam_name,
+ "timestamp": time.time(),
+ "filename": filename
+ })
+
+ except Exception as e:
+ print(f"摄像头 '{cam_name}' 捕获失败: {e}")
+
+ return images
+
+ def get_imu_data(self):
+ """获取IMU传感器数据"""
+ try:
+ imu_data = self.client.getImuData(imu_name="Imu", vehicle_name=self.vehicle_name)
+
+ data = {
+ "timestamp": time.time(),
+ "linear_acceleration": {
+ "x": imu_data.linear_acceleration.x_val,
+ "y": imu_data.linear_acceleration.y_val,
+ "z": imu_data.linear_acceleration.z_val
+ },
+ "angular_velocity": {
+ "x": imu_data.angular_velocity.x_val,
+ "y": imu_data.angular_velocity.y_val,
+ "z": imu_data.angular_velocity.z_val
+ },
+ "orientation": {
+ "w": imu_data.orientation.w_val,
+ "x": imu_data.orientation.x_val,
+ "y": imu_data.orientation.y_val,
+ "z": imu_data.orientation.z_val
+ }
+ }
+
+ self.sensor_data["imu"].append(data)
+ return data
+
+ except Exception as e:
+ print(f"获取IMU数据失败: {e}")
+ return None
+
+ def get_gps_data(self):
+ """获取GPS数据"""
+ try:
+ gps_data = self.client.getGpsData(gps_name="Gps", vehicle_name=self.vehicle_name)
+
+ data = {
+ "timestamp": time.time(),
+ "latitude": gps_data.gnss.geopoint.latitude,
+ "longitude": gps_data.gnss.geopoint.longitude,
+ "altitude": gps_data.gnss.geopoint.altitude,
+ "velocity": {
+ "x": gps_data.gnss.velocity.x_val,
+ "y": gps_data.gnss.velocity.y_val,
+ "z": gps_data.gnss.velocity.z_val
+ }
+ }
+
+ self.sensor_data["gps"].append(data)
+ return data
+
+ except Exception as e:
+ print(f"获取GPS数据失败: {e}")
+ return None
+
+ def manual_control_demo(self, duration=10):
+ """
+ 手动控制演示:前进、转向、停止
+
+ 参数:
+ duration: 演示总时长(秒)
+ """
+ if not self.is_connected or not self.is_api_control_enabled:
+ print("错误: 请先连接并启用API控制")
+ return False
+
+ print(f"\n开始手动控制演示 ({duration}秒)...")
+ print("操作序列: 加速 → 左转 → 右转 → 刹车停止")
+
+ start_time = time.time()
+ sequence = 0
+
+ try:
+ while time.time() - start_time < duration:
+ elapsed = time.time() - start_time
+
+ # 根据时间执行不同控制序列
+ if elapsed < duration * 0.25: # 第一阶段:直线加速
+ controls = airsim.CarControls()
+ controls.throttle = 0.7
+ controls.steering = 0.0
+ self.client.setCarControls(controls, vehicle_name=self.vehicle_name)
+ if sequence < 1:
+ print(" 阶段1: 直线加速")
+ sequence = 1
+
+ elif elapsed < duration * 0.5: # 第二阶段:左转
+ controls = airsim.CarControls()
+ controls.throttle = 0.5
+ controls.steering = 0.3 # 左转
+ self.client.setCarControls(controls, vehicle_name=self.vehicle_name)
+ if sequence < 2:
+ print(" 阶段2: 左转")
+ sequence = 2
+
+ elif elapsed < duration * 0.75: # 第三阶段:右转
+ controls = airsim.CarControls()
+ controls.throttle = 0.5
+ controls.steering = -0.3 # 右转
+ self.client.setCarControls(controls, vehicle_name=self.vehicle_name)
+ if sequence < 3:
+ print(" 阶段3: 右转")
+ sequence = 3
+
+ else: # 第四阶段:减速停止
+ controls = airsim.CarControls()
+ controls.throttle = 0.0
+ controls.brake = 1.0
+ controls.steering = 0.0
+ self.client.setCarControls(controls, vehicle_name=self.vehicle_name)
+ if sequence < 4:
+ print(" 阶段4: 刹车停止")
+ sequence = 4
+
+ # 实时显示车辆状态
+ state = self.get_vehicle_state()
+ if state:
+ print(f"\r速度: {state['speed_kmh']:.1f} km/h | "
+ f"位置: ({state['position']['x']:.1f}, "
+ f"{state['position']['y']:.1f})", end="")
+
+ # 采集传感器数据
+ self.capture_camera_images(["front"]) # 只采集前摄像头
+ self.get_imu_data()
+ self.get_gps_data()
+
+ time.sleep(0.1) # 控制频率 10Hz
+
+ print("\n✓ 手动控制演示完成")
+ return True
+
+ except Exception as e:
+ print(f"\n✗ 控制演示出错: {e}")
+ return False
+
+ def data_collection_demo(self, duration=5):
+ """数据采集演示:采集所有传感器数据"""
+ print(f"\n开始数据采集演示 ({duration}秒)...")
+
+ start_time = time.time()
+ frame_count = 0
+
+ try:
+ while time.time() - start_time < duration:
+ frame_count += 1
+
+ # 采集所有摄像头图像
+ images = self.capture_camera_images()
+
+ # 采集IMU数据
+ imu_data = self.get_imu_data()
+
+ # 采集GPS数据
+ gps_data = self.get_gps_data()
+
+ # 获取车辆状态
+ vehicle_state = self.get_vehicle_state()
+
+ # 记录到日志
+ log_entry = {
+ "frame": frame_count,
+ "timestamp": time.time(),
+ "images": len(images),
+ "imu_data": imu_data is not None,
+ "gps_data": gps_data is not None,
+ "vehicle_state": vehicle_state is not None
+ }
+ self.data_log.append(log_entry)
+
+ print(f"\r采集帧: {frame_count} | "
+ f"图像: {len(images)} | "
+ f"速度: {vehicle_state['speed_kmh'] if vehicle_state else 'N/A':.1f} km/h", end="")
+
+ time.sleep(0.2) # 5Hz采集频率
+
+ print(f"\n✓ 数据采集完成,共采集 {frame_count} 帧")
+ return True
+
+ except Exception as e:
+ print(f"\n✗ 数据采集出错: {e}")
+ return False
+
+ def save_simulation_data(self):
+ """保存所有仿真数据到文件"""
+ try:
+ # 保存车辆状态日志
+ log_file = f"{self.data_dir}/simulation_log.json"
+ with open(log_file, 'w') as f:
+ json.dump(self.data_log, f, indent=2)
+
+ # 保存传感器数据统计
+ stats = {
+ "timestamp": datetime.now().isoformat(),
+ "vehicle_name": self.vehicle_name,
+ "camera_frames": len(self.sensor_data["camera"]),
+ "imu_samples": len(self.sensor_data["imu"]),
+ "gps_samples": len(self.sensor_data["gps"]),
+ "total_log_entries": len(self.data_log)
+ }
+
+ stats_file = f"{self.data_dir}/simulation_stats.json"
+ with open(stats_file, 'w') as f:
+ json.dump(stats, f, indent=2)
+
+ # 生成数据报告
+ report_file = f"{self.data_dir}/report.txt"
+ with open(report_file, 'w') as f:
+ f.write("=" * 50 + "\n")
+ f.write("AirSim无人车仿真数据报告\n")
+ f.write("=" * 50 + "\n\n")
+ f.write(f"仿真时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n")
+ f.write(f"车辆名称: {self.vehicle_name}\n")
+ f.write(f"摄像头帧数: {stats['camera_frames']}\n")
+ f.write(f"IMU采样数: {stats['imu_samples']}\n")
+ f.write(f"GPS采样数: {stats['gps_samples']}\n")
+ f.write(f"日志条目: {stats['total_log_entries']}\n\n")
+ f.write("数据文件:\n")
+ for file in os.listdir(self.data_dir):
+ f.write(f" - {file}\n")
+
+ print(f"\n✓ 仿真数据已保存到: {self.data_dir}")
+ print(f" 日志文件: {log_file}")
+ print(f" 统计数据: {stats_file}")
+ print(f" 报告文件: {report_file}")
+
+ return True
+
+ except Exception as e:
+ print(f"✗ 保存数据失败: {e}")
+ return False
+
+ def run_full_demo(self, control_duration=15, data_duration=8):
+ """运行完整演示"""
+ print("=" * 60)
+ print("AirSimNH 无人车完整仿真演示")
+ print("=" * 60)
+
+ # 步骤1: 连接仿真器
+ if not self.connect():
+ return False
+
+ try:
+ # 步骤2: 启用API控制
+ if not self.enable_api_control(True):
+ return False
+
+ # 步骤3: 手动控制演示
+ if not self.manual_control_demo(control_duration):
+ print("手动控制演示失败,继续其他演示...")
+
+ # 短暂暂停,让车辆完全停止
+ time.sleep(2)
+
+ # 步骤4: 数据采集演示
+ if not self.data_collection_demo(data_duration):
+ print("数据采集演示失败,继续保存数据...")
+
+ # 步骤5: 保存数据
+ self.save_simulation_data()
+
+ return True
+
+ finally:
+ # 步骤6: 清理和退出
+ self.cleanup()
+
+ def cleanup(self):
+ """清理资源"""
+ print("\n正在清理资源...")
+
+ # 停止车辆
+ if self.is_api_control_enabled:
+ controls = airsim.CarControls()
+ controls.brake = 1.0
+ controls.handbrake = True
+ try:
+ self.client.setCarControls(controls, vehicle_name=self.vehicle_name)
+ except:
+ pass
+
+ # 禁用API控制
+ try:
+ self.enable_api_control(False)
+ except:
+ pass
+
+ print("✓ 清理完成")
+
+
+def main():
+ """主函数"""
+ # 创建仿真器实例
+ simulator = AirSimNHCarSimulator(
+ ip="127.0.0.1",
+ port=41451,
+ vehicle_name="PhysXCar"
+ )
+
+ # 运行完整演示
+ try:
+ simulator.run_full_demo(
+ control_duration=20, # 控制演示时长(秒)
+ data_duration=10 # 数据采集时长(秒)
+ )
+
+ print("\n" + "=" * 60)
+ print("仿真演示完成!")
+ print("=" * 60)
+
+ except KeyboardInterrupt:
+ print("\n\n仿真被用户中断")
+ simulator.cleanup()
+ except Exception as e:
+ print(f"\n仿真出错: {e}")
+ simulator.cleanup()
+
+
+if __name__ == "__main__":
+ main()
\ No newline at end of file