A mobile robot only moves horizontally and vertically, navigation avoid obstacle problem

edit

Hi all,

I am simulating a mobile robot that only moves horizontal and vertical, like:

It's length is 7 meters, width is 2.5 meters, and LiDAR is on middle of the robot, the green arrow is the direction it can move, but only one direction at a time.

The wheels are using steering, so it can rotate all wheels for 90 degrees to change direction, like:

In this case, footprint data sets to:[[-1.25, -3.5], [-1.25, 3.5], [1.25, 3.5], [1.25, -3.5]]

In this case, local_planner is: dwa_local_planner

When it moves vertical, it can find the local obstacle, and it can avoid obstacle(Let mobile robot stop), like:

When it moves horizontal, it finds the obstacle, too, but it can't avoid obstacle, like:

Then, I use dynamic_reconfigure update my footprint data to: [[-3.5, -3.5], [-3.5, 3.5], [3.5, 3.5], [3.5, -3.5]]

And, it success, like:

But now, It can't move horizontal to the edge of map, like:

I want it to move like:

But this case it can't avoid obstacle of edge of mobile robot.

My parameters of navigation:

move_base_params.yaml

footprint: [[-1.25, -3.5], [-1.25, 3.5], [1.25, 3.5], [1.25, -3.5]]

controller_frequency: 5.0
controller_patience: 15.0
planner_frequency: 0.0
planner_patience: 5.0
max_planning_retries: 0
oscillation_timeout: 30.0

recovery_behavior_enabled: true
recovery_behaviors: [{name: conservative_reset, type: clear_costmap_recovery/ClearCostmapRecovery}, {name: aggressive_reset, type: clear_costmap_recovery/ClearCostmapRecovery}]
conservative_reset:
  reset_distance: 1.25

dwa_local_planner_params.yaml

base_local_planner: dwa_local_planner/DWAPlannerROS
DWAPlannerROS:

  max_vel_x: 0.5
  min_vel_x: -0.5

  max_vel_y: 0.5
  min_vel_y: -0.5

  max_vel_trans:  0.5
  min_vel_trans:  -0.5

  max_vel_theta: 0.05
  min_vel_theta: -0.05

  acc_lim_x: 5.0
  acc_lim_y: 5.0
  acc_lim_theta: 3.2

  xy_goal_tolerance: 0.03
  yaw_goal_tolerance: 100
  latch_xy_goal_tolerance: true

  sim_time: 1.0
  vx_samples: 10
  vy_samples: 10
  vth_samples: 15

  path_distance_bias: 24.0
  goal_distance_bias: 72.0
  occdist_scale: -1.0           
  forward_point_distance: 0.01  
  stop_time_buffer: 0.2
  scaling_speed: 0.25
  max_scaling_factor: 0.2
  prune_plan: true

  oscillation_reset_dist: 0.001
  oscillation_reset_angle: 0.002

  publish_traj_pc : true
  publish_cost_grid_pc: true

costmap_common_params.yaml

robot_base_frame: base_footprint
transform_tolerance: 0.4
update_frequency: 5.0
publish_frequency: 1.0
obstacle_range: 6.0
publish_voxel_map: true
navigation_map:
  map_topic: /map
obstacles:
  observation_sources: scan
  scan: 
    sensor_frame: laser_link
    data_type: LaserScan 
    topic: scan
    marking: true
    clearing: true

global_costmap_params.yaml

global_costmap:
global_frame: map
static_map: true
raytrace_range: 7.0
resolution: 0.05
z_resolution: 0.2
z_voxels: 10

inflation:
  cost_scaling_factor:  3.0  
  inflation_radius:     0.6 

plugins: 
    - {name: navigation_map, type: "costmap_2d::StaticLayer" }
    - {name: obstacles,  type: "costmap_2d::VoxelLayer" }
    - {name: inflation,  type: "costmap_2d::InflationLayer" }

local_costmap_params.yaml

local_costmap:
  global_frame: odom
  static_map: false
  rolling_window: true
  raytrace_range: 7.0
  resolution: 0.05
  width: 10.0
  height: 10.0
  origin_x: 0.0
  origin_y: 0.0

  inflation:
    cost_scaling_factor: 3.0
    inflation_radius: 3.5

  plugins: 
      - {name: obstacles,  type: "costmap_2d::VoxelLayer" }
      - {name: inflation,  type: "costmap_2d::InflationLayer" }

My questions are:

1. Can I eliminate global costmap obstacle in local costmap?
2. In addition to changing the footprint, how can I change the obstacle inflation in different axis?

3. Is there any other solution to figure out this problem in this ...