为机器人配置传感器#

机器人可以携带传感器进行环境感知。IR-SIM 提供 2D 激光雷达(lidar2d)以及简化的 2D FMCW 激光雷达(fmcw_lidar2d,每条光束附带径向速度),以及可选的视场(FOV)区域;它们都在 YAML 文件中按对象进行配置。本页展示如何配置它们并调节噪声。

LiDAR 配置参数#

下面的 YAML 配置与 Python 脚本展示了带 2D LiDAR 传感器的机器人示例:

import irsim

env = irsim.make()

for i in range(1000):
    env.step()
    env.render(0.05)

    if env.done():
        break

env.end()

与 Python 脚本同名的 YAML 文件:

world:
  height: 10
  width: 10
  step_time: 0.1
  sample_time: 0.1
  offset: [0, 0]
  collision_mode: 'stop'
  control_mode: 'auto'

robot:
  - kinematics: {name: 'diff'}
    shape: {name: 'circle', radius: 0.2}
    state: [1, 1, 0]
    goal: [9, 9, 0]
    behavior: {name: 'dash'}

    sensors:
      - name: 'lidar2d'
        range_min: 0
        range_max: 5
        angle_range: 3.14 
        number: 200
        noise: False
        std: 0.2
        angle_std: 0.2
        offset: [0.1, 0.1, 0.5]
        plot:
          alpha: 0.3
      
obstacle:
  - shape: {name: 'circle', radius: 1.0, center: [1, 1]}
    state: [5, 5, 0]
    unobstructed: True
  
  - shape: {name: 'rectangle', length: 1.5, width: 1.2}
    state: [6, 5, 1]
  
  - shape: {name: 'linestring', vertices: [[10, 5], [4, 0]]}
    state: [0, 0, 0]

  - shape: {name: 'linestring', vertices: [[4, 0], [6, 7]]}
    state: [0, 0, 0]
    unobstructed: True
    plot:
      obj_linestyle: '--'

小技巧

更新顺序

环境会在所有对象完成该时间步移动后再更新传感器,从而避免读数的时间偏差。若手动控制对象,请在 ObjectBase.step(...) 中传入 sensor_step=True,或在更新对象状态后调用 obj.sensor_step()

关键参数说明#

若要配置 2D LiDAR,需在机器人 sensors 部分定义名为 lidar2d 的传感器。主要参数如下:

  • range_min:激光束的最小量程。

  • range_max:激光束的最大量程。

  • angle_range:激光束覆盖的角度范围。完整的 360 度扫描可在 YAML 中使用 6.283185,或在 Python 中使用 2 * pi

  • number:束数量。

  • alpha:激光束透明度。

完整参数列表见 YAML 配置

带噪声的高级 LiDAR 配置#

若需为 LiDAR 添加噪声,可将 noise 设为 True

import irsim

env = irsim.make()

for i in range(1000):
    env.step()
    env.render(0.05)

    if env.done():
        break

env.end()
world:
  height: 10
  width: 10
  step_time: 0.1
  sample_time: 0.1
  offset: [0, 0]
  collision_mode: 'stop'
  control_mode: 'auto'

robot:
  - kinematics: {name: 'diff', noise: True, alpha: [0.1, 0, 0, 0.3]}
    shape: {name: 'circle', radius: 0.2}
    state: [1, 1, 0]
    goal: [9, 9, 0]
    behavior: {name: 'dash', wander: True}

    plot:
      show_trajectory: True

    sensors:
      - type: 'lidar2d'
        range_min: 0
        range_max: 5
        angle_range: 3.14
        number: 200
        noise: True
        std: 0.1
        angle_std: 0.2
        offset: [0, 0, 0]
        alpha: 0.3
      
obstacle:
  - shape: {name: 'circle', radius: 1.0}
    state: [5, 5, 0]

通过 stdangle_std 参数为 LiDAR 添加高斯噪声。其中 std 为距离噪声的标准差,angle_std 为角度噪声的标准差。

FMCW 激光雷达配置参数#

IR-SIM 还提供了一个简化的 2D FMCW 激光雷达传感器,名为 fmcw_lidar2d。它保持与标准 2D 激光雷达相同的光束几何结构,但每条有效光束会额外返回一个标量 radial_velocity(径向速度)测量值。这便于演示多普勒测量如何帮助理解动态障碍物。

下方示例使用一个静止的传感器主体,朝前方覆盖 120 度视场,并设置了多个移动障碍物。开启绘图后,有效回波会按径向速度着色,并在端点处标记出来。

import irsim

env = irsim.make("fmcw_lidar_world.yaml")

for step in range(120):
    env.step()

    scan = env.get_lidar_scan()
    valid_count = int(scan["valid"].sum())
    valid_indices = scan["valid"].nonzero()[0]
    if len(valid_indices) > 0:
        beam_idx = max(valid_indices, key=lambda idx: abs(scan["radial_velocity"][idx]))
        print(
            f"step={step:03d} valid_beams={valid_count:03d} "
            f"beam={beam_idx:03d} range={scan['ranges'][beam_idx]:.3f} "
            f"radial_velocity={scan['radial_velocity'][beam_idx]:.3f}"
        )
    else:
        print(f"step={step:03d} valid_beams={valid_count:03d} no valid returns")

    env.render(0.05, mode="all")

env.end(3)
world:
  height: 10
  width: 10
  step_time: 0.1
  sample_time: 0.1
  offset: [0, 0]
  collision_mode: 'stop'
  control_mode: 'auto'

robot:
  - kinematics: {name: 'diff'}
    shape: {name: 'circle', radius: 0.2}
    state: [2.0, 5.0, 0.0]
    goal: [20.0, 5.0, 0.0]
    vel_min: [-2, -2.0]
    vel_max: [2, 2.0]
    behavior: {name: 'dash'}
    static: true

    sensors:
      - type: 'fmcw_lidar2d'
        range_min: 0.0
        range_max: 8.0
        angle_range: 2.0944   # 120 degrees
        number: 121           # approximately 1 degree per beam, including center beam
        motion_compensate: False
        noise: False
        offset: [0.0, 0.0, 0.0]
        plot:
          alpha: 0.8
          color: 'cyan'
          velocity_linewidth: 2.0
          no_hit_linewidth: 0.25
          no_hit_alpha: 0.01
          show_velocity_markers: True
          velocity_marker_size: 45
          velocity_marker_edge_color: 'black'
          velocity_marker_edge_width: 0.5
          zero_velocity_color: 'cyan'
          positive_velocity_color: 'red'
          negative_velocity_color: 'blue'
          velocity_color_max: 0.6

obstacle:
  - kinematics: {name: 'omni'}
    shape: {name: 'circle', radius: 0.25}
    state: [8.0, 5.0, 0.0]
    goal: [3.0, 5.0, 0.0]
    vel_min: [-0.8, -0.8]
    vel_max: [0.8, 0.8]
    behavior: {name: 'dash'}
    color: 'orangered'
    plot: {show_goal: True}

  - kinematics: {name: 'omni'}
    shape: {name: 'circle', radius: 0.22}
    state: [7.5, 6.8, 0.0]
    goal: [3.2, 4.8, 0.0]
    vel_min: [-0.4, -0.4]
    vel_max: [0.4, 0.4]
    behavior: {name: 'dash'}
    color: 'gold'
    plot: {show_goal: True}

  - kinematics: {name: 'omni'}
    shape: {name: 'circle', radius: 0.20}
    state: [8.6, 3.8, 0.0]
    goal: [3.5, 5.6, 0.0]
    vel_min: [-1.2, -1.2]
    vel_max: [1.2, 1.2]
    behavior: {name: 'dash'}
    color: 'limegreen'
    plot: {show_goal: True}

  - kinematics: {name: 'omni'}
    shape: {name: 'circle', radius: 0.24}
    state: [6.8, 7.6, 0.0]
    goal: [4.6, 6.4, 0.0]
    vel_min: [-0.2, -0.2]
    vel_max: [0.2, 0.2]
    behavior: {name: 'dash'}
    color: 'mediumorchid'
    plot: {show_goal: True}

FMCW 专用关键参数:

  • motion_compensate:是否从测量的径向速度中扣除自身运动分量。

  • velocity_noise_std:施加在 radial_velocity 上的高斯噪声标准差。

可视化参数位于传感器的 plot: 子字典下(与 lidar2d 及对象的 plot: 一致);为保持向后兼容,扁平的顶层键仍然被接受:

  • velocity_color:绘图时是否按径向速度对有效光束着色。

  • velocity_color_max:绘图颜色达到饱和的速度幅值。

返回的扫描数据保留了标准激光雷达的角度元数据,并新增以下字段:

  • radial_velocity:每条光束的标量径向速度。

  • valid:该光束是否在 range_max 范围内有有效回波。

完整可运行示例位于 usage/22fmcw_lidar_world/

视场(FOV)配置参数#

下方的 YAML 配置与 Python 脚本展示了 FOV 中的对象示例。FOV 由对象配置中的 fov(角度)与 fov_radius(半径)参数定义。每个对象都拥有 FOV,可通过 fov_detect_object() 检测视场内的机器人。

import irsim

env = irsim.make()

for i in range(200):
    env.step()

    for obs in env.obstacle_list:
        if obs.fov_detect_object(env.robot):
            print(
                f"The robot is in the FOV of the {obs.name}. The parameters of this obstacle are: state [x, y, theta]: {obs.state.flatten()}, velocity [linear, angular]: {obs.velocity.flatten()}, fov in radian: {obs.fov}."
            )

    env.render(figure_kwargs={"dpi": 100})

env.end()
world:
  height: 50
  width: 50   
  step_time: 0.1 
  sample_time: 0.1  
  offset: [0, 0]  
  collision_mode: 'stop' 
  control_mode: 'auto' 

robot:
  - kinematics: {name: 'diff'}  # omni, diff, acker
    shape: {name: 'circle', radius: 0.4}
    state: [10, 10, 0, 0]
    goal: [45, 45, 0]
    goal_threshold: 0.4
    vel_max: [3, 1]
    vel_min: [-3, -1]
    behavior: {name: 'dash', wander: True, range_low: [15, 15, -3.14], range_high: [35, 35, 3.14]} 
    plot:
        show_goal: True
        show_trajectory: True

obstacle:
  - number: 10
    distribution: {name: 'random', range_low: [10, 10, -3.14], range_high: [40, 40, 3.14]}
    kinematics: {name: 'diff'}
    behavior: {name: 'rvo', vxmax: 1.5, vymax: 1.5, acce: 1.0, factor: 2.0, mode: 'vo', wander: True, range_low: [15, 15, -3.14], range_high: [35, 35, 3.14], target_roles: 'all'}
    vel_max: [3, 3.14]
    vel_min: [-3, -3.14]
    shape:
      - {name: 'circle', radius: 0.5}  # radius
    fov: 1.57 
    fov_radius: 5.0
    plot:
      show_fov: True
      show_arrow: True
      arrow_length: 0.8