添加行人场景生成器

PythonAPI/examples/pedestrian_scenario_generator.py

@@ -0,0 +1,176 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+行人测试场景生成器
+基于已有行人系统生成各种测试场景
+"""
+
+import carla
+import random
+import math
+import time
+import argparse
+
+
+class PedestrianScenarioGenerator:
+    """行人测试场景生成器"""
+
+    def __init__(self, host='localhost', port=2000):
+        self.client = carla.Client(host, port)
+        self.client.set_timeout(10.0)
+        self.world = self.client.get_world()
+        self.pedestrians = []
+        print("[初始化] 行人测试场景生成器已启动")
+
+    def load_walker_blueprints(self):
+        """加载行人蓝图"""
+        blueprint_library = self.world.get_blueprint_library()
+        return blueprint_library.filter('walker.pedestrian.*')
+
+    def spawn_pedestrian(self, transform, behavior='random'):
+        """生成单个行人"""
+        walker_bps = self.load_walker_blueprints()
+        walker_bp = random.choice(walker_bps)
+
+        walker = self.world.spawn_actor(walker_bp, transform)
+
+        # 添加AI控制器
+        controller_bp = self.world.get_blueprint_library().find('controller.ai.walker')
+        controller = self.world.spawn_actor(controller_bp, carla.Transform(), walker)
+
+        self.pedestrians.append({'walker': walker, 'controller': controller})
+        return walker, controller
+
+    def generate_crossing_scenario(self, road_width=10, pedestrian_count=5):
+        """生成过马路测试场景"""
+        print(f"[场景] 生成过马路场景，道路宽度: {road_width}m，行人: {pedestrian_count}")
+
+        # 在道路两侧生成行人
+        for i in range(pedestrian_count):
+            # 随机位置
+            x = random.uniform(-road_width/2, road_width/2)
+            y = random.choice([-5, 5])  # 道路两侧
+
+            transform = carla.Transform(
+                carla.Location(x=x, y=y, z=1.0)
+            )
+
+            walker, controller = self.spawn_pedestrian(transform)
+
+            # 设置过马路目标
+            target = carla.Location(x=x, y=-y, z=1.0)
+            controller.go_to_location(target)
+            controller.set_max_speed(1.4)
+
+            print(f"  行人 #{i+1} 从 ({x:.1f}, {y:.1f}) 到 ({x:.1f}, {-y:.1f})")
+
+    def generate_crowd_scenario(self, center, radius=20, count=20):
+        """生成人群测试场景"""
+        print(f"[场景] 生成人群场景，中心: {center}，半径: {radius}m，人数: {count}")
+
+        for i in range(count):
+            angle = random.uniform(0, 2 * math.pi)
+            distance = random.uniform(0, radius)
+
+            x = center.x + distance * math.cos(angle)
+            y = center.y + distance * math.sin(angle)
+
+            transform = carla.Transform(
+                carla.Location(x=x, y=y, z=1.0)
+            )
+
+            walker, controller = self.spawn_pedestrian(transform)
+
+            # 随机移动目标
+            target_angle = random.uniform(0, 2 * math.pi)
+            target_dist = random.uniform(10, 30)
+            target = carla.Location(
+                x=center.x + target_dist * math.cos(target_angle),
+                y=center.y + target_dist * math.sin(target_angle),
+                z=1.0
+            )
+
+            controller.go_to_location(target)
+            controller.set_max_speed(random.uniform(1.0, 2.0))
+
+    def generate_collision_risk_scenario(self, vehicle_path, pedestrian_count=3):
+        """生成碰撞风险测试场景"""
+        print(f"[场景] 生成碰撞风险场景，行人: {pedestrian_count}")
+
+        # 在车辆路径上生成行人
+        for i in range(pedestrian_count):
+            # 在路径附近随机位置
+            t = random.uniform(0.2, 0.8)  # 路径中间段
+            x = vehicle_path[0].x + (vehicle_path[1].x - vehicle_path[0].x) * t
+            y = vehicle_path[0].y + (vehicle_path[1].y - vehicle_path[0].y) * t
+
+            # 偏移到车道
+            offset = random.uniform(-3, 3)
+            x += offset
+
+            transform = carla.Transform(
+                carla.Location(x=x, y=y, z=1.0)
+            )
+
+            walker, controller = self.spawn_pedestrian(transform)
+
+            # 让行人停留在原地或缓慢移动
+            if random.choice([True, False]):
+                controller.set_max_speed(0.5)  # 缓慢移动
+            else:
+                controller.set_max_speed(0)  # 静止
+
+            print(f"  风险行人 #{i+1} 位置: ({x:.1f}, {y:.1f})")
+
+    def cleanup(self):
+        """清理所有行人"""
+        print(f"[清理] 移除 {len(self.pedestrians)} 个行人")
+        for p in self.pedestrians:
+            if p['controller'].is_alive:
+                p['controller'].stop()
+                p['controller'].destroy()
+            if p['walker'].is_alive:
+                p['walker'].destroy()
+        self.pedestrians.clear()
+
+
+def main():
+    parser = argparse.ArgumentParser(description='行人测试场景生成器')
+    parser.add_argument('--host', default='localhost', help='主机地址')
+    parser.add_argument('--port', type=int, default=2000, help='端口')
+    parser.add_argument('--scenario', default='crossing',
+                       choices=['crossing', 'crowd', 'risk'],
+                       help='场景类型')
+    parser.add_argument('--count', type=int, default=5, help='行人数量')
+    parser.add_argument('--duration', type=int, default=30, help='持续时间')
+
+    args = parser.parse_args()
+
+    generator = PedestrianScenarioGenerator(args.host, args.port)
+
+    try:
+        if args.scenario == 'crossing':
+            generator.generate_crossing_scenario(pedestrian_count=args.count)
+        elif args.scenario == 'crowd':
+            generator.generate_crowd_scenario(
+                carla.Location(x=0, y=0, z=0),
+                count=args.count
+            )
+        elif args.scenario == 'risk':
+            generator.generate_collision_risk_scenario(
+                [carla.Location(x=-50, y=0, z=0), carla.Location(x=50, y=0, z=0)],
+                pedestrian_count=args.count
+            )
+
+        print(f"\n[运行] 场景运行中，持续 {args.duration} 秒...")
+        time.sleep(args.duration)
+
+    except KeyboardInterrupt:
+        print("\n[中断] 用户中断")
+    finally:
+        generator.cleanup()
+        print("[完成] 场景结束")
+
+
+if __name__ == '__main__':
+    main()