版块

• 社区&开源

  社区运营&开源工具

  学习资源 社区运营 一键安装 FishROS2GO
  1.3k
  主题
  4.9k
  帖子

  3

  ls@ls-OMEN-Gaming-Laptop-16-ae0xxx:~/chapter6/chapter6_ws$ ros2 launch fishbot_description gazebo_sim.launch.py
  [INFO] [launch]: All log files can be found below /home/ls/.ros/log/2025-07-14-19-31-02-736177-ls-OMEN-Gaming-Laptop-16-ae0xxx-43738
  [INFO] [launch]: Default logging verbosity is set to INFO
  [INFO] [robot_state_publisher-1]: process started with pid [43741]
  [INFO] [gzserver-2]: process started with pid [43743]
  [INFO] [gzclient-3]: process started with pid [43745]
  [INFO] [spawn_entity.py-4]: process started with pid [43747]
  [robot_state_publisher-1] [INFO] [1752492663.155872124] [robot_state_publisher]: got segment back_caster_link
  [robot_state_publisher-1] [INFO] [1752492663.155987358] [robot_state_publisher]: got segment back_wheel_link
  [robot_state_publisher-1] [INFO] [1752492663.156001201] [robot_state_publisher]: got segment base_footprint
  [robot_state_publisher-1] [INFO] [1752492663.156010853] [robot_state_publisher]: got segment base_link
  [robot_state_publisher-1] [INFO] [1752492663.156019436] [robot_state_publisher]: got segment camera_link
  [robot_state_publisher-1] [INFO] [1752492663.156027798] [robot_state_publisher]: got segment front_caster_link
  [robot_state_publisher-1] [INFO] [1752492663.156036241] [robot_state_publisher]: got segment imu_link
  [robot_state_publisher-1] [INFO] [1752492663.156044576] [robot_state_publisher]: got segment laser_cylinder_link
  [robot_state_publisher-1] [INFO] [1752492663.156052841] [robot_state_publisher]: got segment laser_link
  [robot_state_publisher-1] [INFO] [1752492663.156060997] [robot_state_publisher]: got segment left_wheel_link
  [gzclient-3] Gazebo multi-robot simulator, version 11.10.2
  [gzclient-3] Copyright (C) 2012 Open Source Robotics Foundation.
  [gzclient-3] Released under the Apache 2 License.
  [gzclient-3] http://gazebosim.org
  [gzclient-3]
  [spawn_entity.py-4] Traceback (most recent call last):
  [spawn_entity.py-4] File "/opt/ros/humble/lib/gazebo_ros/spawn_entity.py", line 31, in <module>
  [spawn_entity.py-4] from lxml import etree as ElementTree
  [spawn_entity.py-4] ModuleNotFoundError: No module named 'lxml'
  [gzserver-2] Gazebo multi-robot simulator, version 11.10.2
  [gzserver-2] Copyright (C) 2012 Open Source Robotics Foundation.
  [gzserver-2] Released under the Apache 2 License.
  [gzserver-2] http://gazebosim.org
  [gzserver-2]
  [ERROR] [spawn_entity.py-4]: process has died [pid 43747, exit code 1, cmd '/opt/ros/humble/lib/gazebo_ros/spawn_entity.py -topic /robot_description -entity fishbot --ros-args'].
  [gzserver-2] [Wrn] [gazebo_ros_init.cpp:178]
  [gzserver-2] # # ####### ####### ### ##### #######
  [gzserver-2] ## # # # # # # # #
  [gzserver-2] # # # # # # # # #
  [gzserver-2] # # # # # # # # #####
  [gzserver-2] # # # # # # # # #
  [gzserver-2] # ## # # # # # # #
  [gzserver-2] # # ####### # ### ##### #######
  [gzserver-2]
  [gzserver-2] This version of Gazebo, now called Gazebo classic, reaches end-of-life
  [gzserver-2] in January 2025. Users are highly encouraged to migrate to the new Gazebo
  [gzserver-2] using our migration guides (https://gazebosim.org/docs/latest/gazebo_classic_migration?utm_source=gazebo_ros_pkgs&utm_medium=cli)
  [gzserver-2]
  [gzserver-2]
  [gzclient-3] [Msg] Waiting for master.
  [gzclient-3] [Msg] Connected to gazebo master @ http://127.0.0.1:11345
  [gzclient-3] [Msg] Publicized address: 10.197.145.93
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/Optical_flow"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/Sunflower"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/TensorRT"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/Test"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/YOLO"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/cartographer"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/define"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/final_reference"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/lidarteachws"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/omniverse"
  [gzclient-3] [Err] [InsertModelWidget.cc:403] Missing model.config for model "/home/ls/Desktop/robotcontrol"
  [gzclient-3] [Wrn] [Event.cc:61] Warning: Deleting a connection right after creation. Make sure to save the ConnectionPtr from a Connect call
  [gzclient-3] [Wrn] [Publisher.cc:135] Queue limit reached for topic /gazebo/default/user_camera/pose, deleting message. This warning is printed only once.
  ^C[WARNING] [launch]: user interrupted with ctrl-c (SIGINT)
  [gzclient-3] [Wrn] [Event.cc:61] Warning: Deleting a connection right after creation. Make sure to save the ConnectionPtr from a Connect call
  [robot_state_publisher-1] [INFO] [1752492883.290270574] [rclcpp]: signal_handler(signum=2)
  [gzserver-2] [Msg] Waiting for master.
  [gzserver-2] [Msg] Connected to gazebo master @ http://127.0.0.1:11345
  [gzserver-2] [Msg] Publicized address: 10.197.145.93
  [gzserver-2] [Msg] Loading world file [/home/ls/chapter6/chapter6_ws/install/fishbot_description/share/fishbot_description/world/custom_room.world]
  [INFO] [robot_state_publisher-1]: process has finished cleanly [pid 43741]
  [INFO] [gzserver-2]: process has finished cleanly [pid 43743]
  [gzserver-2]
  [gzserver-2]
  [INFO] [gzclient-3]: process has finished cleanly [pid 43745]
  [gzclient-3]
  [gzclient-3]

• 产品&店铺

  鱼香ROS 店铺&产品相关问题

  FishBot二驱机器人 FishBot四驱机器人V2 多协议传输控制器
  446
  主题
  2.9k
  帖子

  L

  通过网盘分享的文件：FishBot_VMware
  链接: https://pan.baidu.com/s/1tVDm_pWtZ0m1XENqSAKLUg?pwd=fish 提取码: fish
  --来自百度网盘超级会员v2的分享

  1、下载网盘里的文件
  a7bb14e0-6876-44a9-8f4c-85dbf659b4ad-image.png

  2、打开VMware-workstation-full-17.5.2-23775571.exe，一直点击下一步，安装VMware。在最后一步完成时输入密钥，密钥为VM全版本许可证密钥里的VMware-workstation 17的密钥。
  安装完成之后桌面上会有如下图标
  14992647-5f13-4378-9109-77fc5f6bdfbe-image.png

  3、解压网盘里的镜像文件压缩包
  095bca56-ce46-486a-9391-011741d79a3e-image.png

  4、双击点开VMware，在文件里面打开解压得到的Ubuntu22_FishBot
  0b899981-1ccb-407c-a92d-408630e63035-image.png
  e3abc024-40da-49e5-808a-eff512aeda56-image.png

  5、打开之后会弹出导入虚拟机的弹窗，给虚拟机起一个名字，安排合适的路径
  5fa6ea21-667e-4f4d-b2ff-39312bace706-image.png

  设置完之后点击导入

  6、导入之后，在编辑虚拟机设置里面可以修改内存大小，直接点击“开启此虚拟机”进入fishros的虚拟机，密码ros
  59a3a74d-ec21-460c-a176-21b54bb289d6-image.png

• VIP问答专区

  针对社区VIP用户问答专区

  66
  主题
  259
  帖子

  2

  e1d744a8-c2f1-4092-a3c9-63c5000e7532-ccadef6b-38e3-4d46-a578-a643ddc21c77.png

  下面是我的代码
  import launch
  import launch_ros
  from ament_index_python.packages import get_package_share_directory
  import os
  import launch_ros.parameter_descriptions

  def generate_launch_description():
  urdf_package_path=get_package_share_directory('fishbot_description')
  default_urdf_path=os.path.join(urdf_package_path,'urdf','first_robot.urdf')

  action_declare_arg_mode_path=launch.actions.DeclareLaunchArgument( name='model',default_value=str(default_urdf_path),description='加载的模型文件路径' ) substitutions_result=launch.substitutions.Command(['cat ',launch.substitutions.launch_configuration('model')]) robot_description_value=launch_ros.parameter_descriptions.ParameterValue(substitutions_result,value_type=str) action_robot_state_publisher=launch_ros.actions.Node( package='robot_state_publisher', executable='robot_state_publisher', parameters=[{'robot_description':robot_description_value}] ) action_joint_state_publisher=launch_ros.actions.Node( package='joint_state_publisher', executable='joint_state_publisher', ) action_rviz_node=launch_ros.actions.Node( package='rviz2', executable='rviz2', ) return launch.launch_description([ action_declare_arg_mode_path, action_robot_state_publisher, action_joint_state_publisher, action_rviz_node, ])
• ROS2机器人开发：从入门到实践

  《ROS2机器人开发：从入门到实践》书籍配套资料答疑专区

  ROS 2相关问题
  1.1k
  主题
  4.3k
  帖子

  Q

  安装mirco-ROS-Agent，colcon build时，出现如下错误

  7fcd7191-e5ec-49f7-ad18-e878624fcd45-image.png

• 综合问题

  综合问题讨论版块 | 版主 @吴凯荣 @小鱼

  1.0k
  主题
  3.6k
  帖子

  秋

  我的系统是ubuntu20.04，用的ros是galactic,我在同步之前录制了一个bag包，包含6个相机和一个激光雷达的信息系，下面是bag包的相关信息：

  c606@ubuntu:~/fyx/data_bag$ ros2 bag info rosbag2_2025_07_10-03_25_26 Files: rosbag2_2025_07_10-03_25_26_0.db3 Bag size: 2.4 GiB Storage id: sqlite3 Duration: 17.252s Start: Jul 10 2025 03:25:28.197 (1752143128.197) End: Jul 10 2025 03:25:45.450 (1752143145.450) Messages: 999 Topic information: Topic: /rslidar_points | Type: sensor_msgs/msg/PointCloud2 | Count: 173 | Serialization Format: cdr Topic: /camera1/pylon_ros2_camera_node/image_raw | Type: sensor_msgs/msg/Image | Count: 135 | Serialization Format: cdr Topic: /camera3/pylon_ros2_camera_node/image_raw | Type: sensor_msgs/msg/Image | Count: 137 | Serialization Format: cdr Topic: /camera2/pylon_ros2_camera_node/image_raw | Type: sensor_msgs/msg/Image | Count: 123 | Serialization Format: cdr Topic: /camera4/pylon_ros2_camera_node/image_raw | Type: sensor_msgs/msg/Image | Count: 144 | Serialization Format: cdr Topic: /camera5/pylon_ros2_camera_node/image_raw | Type: sensor_msgs/msg/Image | Count: 146 | Serialization Format: cdr Topic: /camera6/pylon_ros2_camera_node/image_raw | Type: sensor_msgs/msg/Image | Count: 141 | Serialization Format: cdr

  接着我写了一个功能包，作用是接受相机和雷达的话题，将话题内容同步后再以话题的形式发送出去，在写好节点代码后，我先通过命令ros2 run sys_time sys_time_node开启节点，再通过ros2 bag record -o synced_bag /clock $(ros2 topic list | grep "synced")命令开启录制，最后通过ros2 bag play rosbag2_2025_07_10-03_25_26 --loop --clock --read-ahead-queue-size 10000命令播放原始bag包。

  在播放bag包后我发现同步节点的回调函数没有被调用，结束录制命令后我通过rqt_bag synced_bag命令查看时间戳，发现没有时间戳，这证明同步未进行或者同步失败了。

  下面是我同步节点的运行日志：
  c606@ubuntu:~/fyx/e2e_ws$ ros2 run sys_time sys_time_node
  [INFO] [1752477350.845923197] [multi_sensor_sync]: 🚀 多传感器同步节点已启动 (ApproximateTime策略)
  [WARN] [1752477359.044116748] []: Messages of type 6 arrived out of order (will print only once)
  [WARN] [1752477376.810937820] []: Messages of type 3 arrived out of order (will print only once)
  [WARN] [1752477376.825926258] []: Messages of type 0 arrived out of order (will print only once)
  [WARN] [1752477376.831166002] []: Messages of type 2 arrived out of order (will print only once)
  [WARN] [1752477376.840478915] []: Messages of type 5 arrived out of order (will print only once)
  [WARN] [1752477376.891147241] []: Messages of type 1 arrived out of order (will print only once)
  [WARN] [1752477377.102514764] []: Messages of type 4 arrived out of order (will print only once)

  下面是我的同步节点代码：

  #include <rclcpp/rclcpp.hpp> #include <sensor_msgs/msg/image.hpp> #include <sensor_msgs/msg/point_cloud2.hpp> #include <message_filters/subscriber.h> #include <message_filters/synchronizer.h> #include <message_filters/sync_policies/approximate_time.h> using ImageMsg = sensor_msgs::msg::Image; using PointCloud2Msg = sensor_msgs::msg::PointCloud2; class MultiSensorSync : public rclcpp::Node { public: MultiSensorSync() : Node("multi_sensor_sync") { // 增强QoS配置 auto qos = rclcpp::SensorDataQoS() .keep_last(20) // 增大缓存队列 .reliable() // 可靠传输模式 .durability_volatile(); // 允许历史数据 initPublishers(qos); initSubscribersAndSync(qos); RCLCPP_INFO(this->get_logger(), "🚀 多传感器同步节点已启动 (ApproximateTime策略)"); } private: void initPublishers(const rclcpp::QoS& qos) { for (int i = 0; i < 6; ++i) { cam_pubs_.push_back( this->create_publisher<ImageMsg>("/camera" + std::to_string(i+1) + "/synced/image_raw", qos) ); } lidar_pub_ = this->create_publisher<PointCloud2Msg>("/rslidar_points/synced", qos); } void initSubscribersAndSync(const rclcpp::QoS& qos) { // 创建相机订阅器 for (int i = 0; i < 6; ++i) { cam_subs_.push_back( std::make_shared<message_filters::Subscriber<ImageMsg>>( this, "/camera" + std::to_string(i+1) + "/pylon_ros2_camera_node/image_raw", qos.get_rmw_qos_profile() ) ); } // 创建雷达订阅器 lidar_sub_ = std::make_shared<message_filters::Subscriber<PointCloud2Msg>>( this, "/rslidar_points", qos.get_rmw_qos_profile() ); // 配置ApproximateTime策略 using SyncPolicy = message_filters::sync_policies::ApproximateTime< ImageMsg, ImageMsg, ImageMsg, ImageMsg, ImageMsg, ImageMsg, PointCloud2Msg>; // 创建同步器（增大时间窗口） sync_ = std::make_shared<message_filters::Synchronizer<SyncPolicy>>( SyncPolicy(50), // 队列大小增至50 *cam_subs_[0], *cam_subs_[1], *cam_subs_[2], *cam_subs_[3], *cam_subs_[4], *cam_subs_[5], *lidar_sub_ ); // 注册回调并绑定生命周期 sync_->registerCallback( std::bind(&MultiSensorSync::syncCallback, this, std::placeholders::_1, std::placeholders::_2, std::placeholders::_3, std::placeholders::_4, std::placeholders::_5, std::placeholders::_6, std::placeholders::_7 ) ); } void syncCallback( const ImageMsg::ConstSharedPtr& cam1, const ImageMsg::ConstSharedPtr& cam2, const ImageMsg::ConstSharedPtr& cam3, const ImageMsg::ConstSharedPtr& cam4, const ImageMsg::ConstSharedPtr& cam5, const ImageMsg::ConstSharedPtr& cam6, const PointCloud2Msg::ConstSharedPtr& lidar) { // 使用第一个相机的时间戳作为基准 //const auto& base_stamp = cam1->header.stamp; const auto& base_stamp = lidar->header.stamp; // 发布所有相机数据（统一时间戳） publishWithStamp(cam_pubs_[0], cam1, base_stamp); publishWithStamp(cam_pubs_[1], cam2, base_stamp); publishWithStamp(cam_pubs_[2], cam3, base_stamp); publishWithStamp(cam_pubs_[3], cam4, base_stamp); publishWithStamp(cam_pubs_[4], cam5, base_stamp); publishWithStamp(cam_pubs_[5], cam6, base_stamp); // 发布雷达数据（统一时间戳） auto new_lidar = std::make_shared<PointCloud2Msg>(*lidar); new_lidar->header.stamp = base_stamp; lidar_pub_->publish(*new_lidar); // 修复时间戳访问方式 RCLCPP_INFO(this->get_logger(), "🔥 同步发布 @ %d.%09d", base_stamp.sec, base_stamp.nanosec); } void publishWithStamp(rclcpp::Publisher<ImageMsg>::SharedPtr pub, const ImageMsg::ConstSharedPtr& msg, const builtin_interfaces::msg::Time& stamp) { auto new_msg = std::make_shared<ImageMsg>(*msg); new_msg->header.stamp = stamp; pub->publish(*new_msg); } // 成员变量 std::vector<rclcpp::Publisher<ImageMsg>::SharedPtr> cam_pubs_; std::vector<std::shared_ptr<message_filters::Subscriber<ImageMsg>>> cam_subs_; rclcpp::Publisher<PointCloud2Msg>::SharedPtr lidar_pub_; std::shared_ptr<message_filters::Subscriber<PointCloud2Msg>> lidar_sub_; using SyncPolicy = message_filters::sync_policies::ApproximateTime< ImageMsg, ImageMsg, ImageMsg, ImageMsg, ImageMsg, ImageMsg, PointCloud2Msg>; std::shared_ptr<message_filters::Synchronizer<SyncPolicy>> sync_; }; int main(int argc, char** argv) { rclcpp::init(argc, argv); rclcpp::executors::MultiThreadedExecutor executor; auto node = std::make_shared<MultiSensorSync>(); executor.add_node(node); executor.spin(); rclcpp::shutdown(); return 0; }

  最后还有一点，我录制的原始bag中相机和雷达的频率在10hz左右，下面是原始bag中的时间戳截图：
  Screenshot from 2025-07-14 00-46-07.png

• 机械臂力控板块

  机械臂力控板块 | 版主： @陆门马

  5
  主题
  11
  帖子

  2

  400f79ff-8a2f-4584-b099-ffe7dbc31462.jpeg

  机械臂，遥操作+加力反馈，如何实现？

  遥操作：通过手柄，手机，或者任何形式的“遥控器”遥控机械臂的运动。最简单的形式就是以一个机械臂（主臂）遥控另一个机械臂（从臂）。

  力反馈：当从臂碰到障碍物时，主臂可以以力的形式感受到这个障碍物，所谓“感受”就是操作者感受到主臂的力。

  代码：https://gitee.com/qingqing-gaq/feel_force.git

  视频：https://www.bilibili.com/video/BV1KoEqzyEFR/?vd_source=36451ca805358187c4a0efa1c5694b35

  我的实现思路：
  软件：rso2
  硬件：双通道usb转can模块，六个达妙4310电机。

  六个电机，组成两个三自由度机械臂。一个can通道控制一个机械臂。

  大体逻辑：
  主臂时重力补偿模式，因为只有重力补偿才能任意拖动。主臂把实时的关节位置，发送给从臂，因此从臂必然是位置控制模式，这样才能跟踪主臂的位置。
  那么主臂如何感受从臂的力呢？这里就需要电机有力矩接口了（达妙电机的mit模式），当从臂电机受到阻力时，pid控制力就会变大，电机反馈会的力也会变大，这个变大的值就是阻力了。把这个力反馈给主臂即可。
  程序运行逻辑：
  23467808-ec76-44af-b4b5-f339ae8f85d4.png
  操作逻辑的代码：

  std::cout << "JointPositionControl::execute..............." << std::endl; //从对应的can接收线程中读取电机状态，并更新到机械臂类储存关节状态的变量（下面会用到）中 arm->update_current_joint_states();//一定要在get_current_joint_states()之前执行 arm2->update_current_joint_states();//一定要在get_current_joint_states()之前执行 //从机械臂类里返回储存关节状态的变量 const auto &curret_joint_states = arm->get_current_joint_states(); const auto &curret_joint_states2 = arm2->get_current_joint_states(); //根据当前的关节状态，计算逆动力学，进行重力补偿。 arm->computeInverseDynamics(arm->current_joint_positions, arm->current_joint_velocities, arm->current_joint_acceleration, arm->gravity_joint_tauqes); arm2->computeInverseDynamics(arm2->current_joint_positions, arm2->current_joint_velocities, arm2->current_joint_acceleration, arm2->gravity_joint_tauqes); std::vector<float> pos1(arm->getDof(), 0.0f); std::vector<float> pos2(arm->getDof(), 0.0f); std::vector<float> vel(arm->getDof(), 0.0f); std::vector<float> vel2(arm->getDof(), 0.0f); std::vector<float> tor(arm->getDof(), 0.0f); std::vector<float> tor2(arm2->getDof(), 0.0f); for (size_t i = 0; i < arm->getDof(); i++) { if(i==2) {//三号电机转向和模型相反，所以要单独加一个负号 tor[i] = -arm->gravity_joint_tauqes(i);//发送给主臂的重力补偿力矩 tor2[i] = -arm2->gravity_joint_tauqes(i);//发送给从臂的重力补偿力矩 } else { tor[i] = arm->gravity_joint_tauqes(i);//发送给主臂的重力补偿力矩 tor2[i] = arm2->gravity_joint_tauqes(i);//发送给主臂的重力补偿力矩 } //主臂的关节速度，发送给从臂 vel2[i] = 0.10 * curret_joint_states.at(i).velocity; // 系数越大手感越硬。 //主臂的关节位置，发送给从臂 pos2[i] = curret_joint_states.at(i).position; } float torque_threshold = 0.04f; // 设置阈值 //从臂受到的阻力，返回给主臂 for (size_t i = 0; i < arm->getDof(); i++) { if (std::abs(curret_joint_states2.at(i).torque) > torque_threshold) { tor[i] += -0.5 * (curret_joint_states2.at(i).torque - tor2[i]); // 如果超过阈值，按原逻辑计算 } else { tor[i] += 0.0f; // 否则赋值为 0 } } // pos2[0] = 0; tor2[0] = 0; arm->send_joint_states(pos1, vel, tor);//向主臂发送关节的状态 arm2->send_joint_states(pos2, vel2, tor2);//向从臂发送关节的状态

  其中：主臂位置发送给从臂的逻辑好理解。

  pos2[i] = curret_joint_states.at(i).position; //curret_joint_states.at(i).position：主臂的实时关节位置 //pos2[i]：发送给从臂的关节位置

  这一句是从臂的力矩反馈给主臂。有一步操作是从臂反馈的力矩减去了重力补偿力矩，这是因为，反馈的电机力矩是pid产生的。这个pid产生的控制力矩就包含重力力矩，如果把这个力也算上发送给主臂，那主臂就会变的很称，因为从臂把重力也传过来了。

  tor[i] += -0.5 * (curret_joint_states2.at(i).torque - tor2[i]); //tor[i]： 发送给主臂的关节力矩 //curret_joint_states2.at(i).torque：从臂电机反馈会来的力矩 //tor2[i]：发送给从臂的重力补偿力矩。 //-0.5：缩放系数

  这个if判断的作用是，避免从臂传来的力一直在小幅度波动，只有从臂反馈的力大于一个值，这个力才会被加载到主臂上。

  float torque_threshold = 0.04f; // 设置阈值 //从臂受到的阻力，返回给主臂 for (size_t i = 0; i < arm->getDof(); i++) { if (std::abs(curret_joint_states2.at(i).torque) > torque_threshold) { tor[i] += -0.5 * (curret_joint_states2.at(i).torque - tor2[i]); // 如果超过阈值，按原逻辑计算 } else { tor[i] += 0.0f; // 否则赋值为 0 } }

  这又是一个神奇的操作。就是让从臂有期望速度。作用就是会影响操作手感。

  vel2[i] = 0.10 * curret_joint_states.at(i).velocity; // 系数越大手感越硬。

  ok这就是整个的实现逻辑，非常简陋。

  程序框架简介：https://zwf9l1z0bm3.feishu.cn/docx/Ee3EdW8meoy79uxzBQVcXe11nfh?from=from_copylink
  qq交流群：523255719

• 历史归档板块

  移动机器人制作 机器人相关领域 机械臂运动规划
  358
  主题
  1.6k
  帖子

  3

  @B21022328 解决了吗