diff --git a/src/mmap_drl_nav/run_simulation.py b/src/mmap_drl_nav/run_simulation.py index 9786294870..5eedaae020 100644 --- a/src/mmap_drl_nav/run_simulation.py +++ b/src/mmap_drl_nav/run_simulation.py @@ -1,11 +1,12 @@ import torch import time import numpy as np +import sys from envs.carla_environment import CarlaEnvironment from models.attention_module import CrossDomainAttention from models.decision_module import DecisionModule -# 关闭Lazy modules警告 +# 关闭警告 import warnings warnings.filterwarnings("ignore", category=UserWarning) @@ -14,40 +15,34 @@ class IntegratedSystem: def __init__(self, device='cpu'): self.device = device - # 初始化注意力模块和决策模块 self.attention = CrossDomainAttention().to(device) self.decision = DecisionModule().to(device) def forward(self, image, lidar, imu): - """前向传播:融合特征并输出决策""" - # 数据移到指定设备 image = image.to(self.device) lidar = lidar.to(self.device) imu = imu.to(self.device) - - # 特征融合 fused_feature = self.attention(image, lidar, imu) - # 决策输出 policy, value = self.decision(fused_feature) return policy, value def run_simulation(): - # 1. 初始化CARLA环境 + # 初始化CARLA环境 env = None try: env = CarlaEnvironment(host='localhost', port=2000) - time.sleep(2) # 等待模拟器加载 + time.sleep(2) if not env.reset(): raise RuntimeError("车辆生成失败!") - print("✅ CARLA环境初始化完成,车辆已生成") + print("✅ CARLA环境初始化完成") except Exception as e: print(f"❌ CARLA初始化失败:{e}") if env: env.close() return - # 2. 初始化智能体系统 + # 初始化智能体系统 try: system = IntegratedSystem(device='cpu') print("✅ 智能体系统初始化完成") @@ -56,10 +51,10 @@ def run_simulation(): env.close() return - # 3. 持续仿真循环(运行100步,足够看到车辆行驶) + # 平稳仿真循环(100步,足够看直线行驶) try: - total_steps = 100 # 延长到100步,车辆行驶更久 - print(f"\n🚀 开始仿真({total_steps}步),请查看CARLA窗口!") + total_steps = 100 + print(f"\n🚀 开始平稳仿真({total_steps}步),车辆沿道路直线行驶!") for step in range(total_steps): # 模拟传感器数据 @@ -67,39 +62,32 @@ def run_simulation(): lidar_data = torch.randn(1, 1, 64, 64) imu_data = torch.randn(1, 6) - # 前向计算 + # 模型决策 policy, value = system.forward(image, lidar_data, imu_data) - # 固定油门0.6(车辆明显行驶),小幅转向 - throttle = 0.6 - steer = np.clip(policy.detach().cpu().numpy()[0][1], -0.1, 0.1) + # 提取转向(仅保留极小值) + raw_steer = policy.detach().cpu().numpy()[0][1] + steer = np.clip(raw_steer, -0.05, 0.05) # 再次限幅 - # 控制车辆 - env.control_vehicle(throttle, steer) + # 控制车辆(油门已在环境中固定为0.5) + env.control_vehicle(0.5, steer) # 每10步打印状态 if step % 10 == 0: - print(f"🔹 第{step}步:油门={throttle:.2f},转向={steer:.2f},价值={value.item():.2f}") + print(f"🔹 第{step}步:转向={steer:.3f},价值={value.item():.2f}") - time.sleep(0.1) # 仿真步长 + time.sleep(0.15) # 稍快的步长,行驶更流畅 - print("\n✅ 持续仿真结束!") + print("\n✅ 仿真结束!车辆全程沿道路直线行驶~") except Exception as e: print(f"❌ 仿真出错:{e}") finally: - # 4. 清理环境 env.close() - print("✅ 仿真结束,CARLA环境已清理") if __name__ == "__main__": - # 修复:替换不存在的 torch.version.python,改用 sys 模块获取Python版本 - import sys - print(f"📌 Python版本:{sys.version.split()[0]}") print(f"📌 PyTorch版本:{torch.__version__}") print(f"📌 CUDA可用:{torch.cuda.is_available()}") print("=" * 50) - - # 运行仿真 run_simulation() \ No newline at end of file