mr-tony921 / xju-robot Goto Github PK

The problem of incompatibility of the noetic version before this step has been solved. However, when I tried point-to-point navigation, I found that the navigation could not be performed after a given target point, and the planned route could not appear. I have uninstalled the navigation package. I want to know whether this is related to the ros version?

看完了博主使用二维码辅助定位实现自动回充的视频，感觉技术上是可行的，不过在现实工作场景中，由于光线和二维码可能被遮挡等问题，效果可能会打折扣。现在有通过激光去识别充电装上的特殊结构（如V形槽）来进行辅助定位和回充的，不知道博主是否有了解？

The ackermann control in your code servo_command.py means:

1. the drive wheels share the same velocity
2. the front wheels share the same steering angle
   I think that's enough for most low speed circumstance.
   Yet for a more accurate model, you can modify it like
   accurate

I've added nonlinear factor in case of 0 throttle input with non-0 output (like diff-drive vehicle), the code is like below:

throttle = data.drive.speed
steer = data.drive.steering_angle # angular velocity in fact

if abs(steer) < steer_threshold:
        pub_vel_left_rear_wheel.publish(throttle * 31.75)
        pub_vel_right_rear_wheel.publish(throttle * 31.75)
        pub_pos_left_steering_hinge.publish(steer)
        pub_pos_right_steering_hinge.publish(steer)

        rospy.loginfo("throttle(target velocity): %f", throttle* 31.75)
        rospy.loginfo("steer(target angular velocity): %f", steer)
    elif ( abs(throttle) > abs(steer * d)):
        throttle_left = (throttle - steer * d) * 31.75
        throttle_right = (throttle + steer * d) * 31.75
        R = throttle/steer
        if (R < 0):
            steer_left = math.atan2(l, R + d) - math.pi
            steer_right = math.atan2(l, R - d) - math.pi
        else:
            steer_left  = math.atan2(l, R - d)
            steer_right = math.atan2(l, R + d)
        pub_vel_left_rear_wheel.publish(throttle_left)
        pub_vel_right_rear_wheel.publish(throttle_right)
        pub_pos_left_steering_hinge.publish(steer_left)
        pub_pos_right_steering_hinge.publish(steer_right)

        rospy.loginfo("throttle_left: %f", throttle_left)
        rospy.loginfo("throttle_right: %f", throttle_right)
        rospy.loginfo("steer_left: %f", steer_left)
        rospy.loginfo("steer_right: %f", steer_right)
    else:
        pub_vel_left_rear_wheel.publish(throttle * 31.75)
        pub_vel_right_rear_wheel.publish(throttle * 31.75)
        pub_pos_left_steering_hinge.publish(steer)
        pub_pos_right_steering_hinge.publish(steer)

        rospy.loginfo("throttle(target velocity): %f", throttle* 31.75)
        rospy.loginfo("steer(target angular velocity): %f", steer)

I've tested it with teleop tools in gazebo, it seems working well.

By the way, I wonder why there's a 31.75 product factor in throttle.

I would be very appreciating your consideration of testing this model.

下载了原工程后，没有改任何代码
输入：roslaunch xju_pnc move_base_flex.launch
启动该launch后，设置障碍物和目标点

发现小车直接撞过去了，如何解决呢

I have forked this repo and modified for noetic. Can you create a new noetic branch so I can create a pull request?

您好，请问该仿真环境的所有代码您是已经全部上传以及开源吗？我是否可以使用您的仿真环境进行相应的路径规划算法相关的研究工作呢？具体就是我基于A*算法进行了一些改进工作，已经在MATLAB环境下进行算法的验证了，现在想着可以在ros环境下在次进行更加符合机器人工作的一些研究。

lio-sam好像缺少了，在thirdparty目录下没看到

你好，在文件/xju-robot/src/pnc/node/mbf_bridge_node.cpp中，导入了头文件#include "costmap_2d/keepoutzone.h'，但是在costmap_2d文件中，并没有找到keepoutzone.h。

fatal error: costmap_2d/keepOutZone.h: 没有那个文件或目录
7 | #include "costmap_2d/keepOutZone.h"
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.

环境是Ubuntu18.04 ros melodic
大佬求帮忙看一下

请问能完整工程能提供下么？包括LIO-sam等

您好，XJU移动机器人仿真-第8期 禁行区/虚拟墙用到的插件keepout Layer能不能上传一下，src目录下似乎没有

../xju-robot/devel/lib/mbf_costmap_nav/mbf_costmap_nav: symbol lookup error: .../xju-robot/devel/lib/liblayers.so: undefined symbol: ZN3tf212getTimestampIN13geometry_msgs13PointStamped_ISaIvEEEEERKN3ros4TimeERKT
[move_base_flex-4] process has died [pid 7761, exit code 127, cmd .../xju-robot/devel/lib/mbf_costmap_nav/mbf_costmap_nav cmd_vel:=/cmd_vel odom:=odom __name:=move_base_flex

Seems like DLu/navigation_layers#69 and ros-planning/navigation#938

参考您之前的工程，hdl_localization用做定位并没有使用odom而是用lidar_link连接整棵tf树，但是我在运行过程中却报了错，整棵树是不完整的，lidar_link也是不存在的，我找了整个工程文件似乎没发现他在什么地方定义了，请问这个lidar_link是怎么配置的呢？感谢

Hello, I don't know why the lidar scan clearly has data, but the cost map is not inflated, and the car will hit the obstacle when planning the path.

麻烦帮忙看下这个错误，装了很多次都不行

启动relocation.cpp 文件后，不能像b站中大佬的效果，基本上没有起作用，确保节点编译通过、能订阅地图和雷达的话题，如下输出给定的就是最后计算出的，好像没有起作用 ：

[INFO] [1680866422.659020978] [range_relocatoin]: Receive original initial pose for amcl node [0.003, 0.217, 0.399]
[INFO] [1680866423.726686561] [range_relocatoin]: Get new best pose [0.003, 0.217, 0.399] score [279], time go 1.068

[INFO] [1680866413.634283631] [range_relocatoin]: Receive original initial pose for amcl node [-0.213, 0.280, -0.327]
[INFO] [1680866414.578891224] [range_relocatoin]: Get new best pose [-0.213, 0.280, -0.327] score [49], time go 0.945