Markov Localization Explained
In this tutorial, I explain the math and theory of robot localization and I will solve an example of Markov localization.
ADAS
Car Detection Using Single Shot MultiBox Detector (SSD Convolutional Neural Network) in ROS Using Caffe
This work is similar to the previous work here, but this time I used Single Shot MultiBox Detector (SSD) for car detection. Installation is similar, clone the SSD Caffe:
|
1 2 3 |
git clone https://github.com/weiliu89/caffe.git ssdcaffe cd ssdcaffe git checkout 5365d0dccacd18e65f10e840eab28eb65ce0cda |
add the following lines to your Makefile.config
|
1 2 |
INCLUDE_DIRS += /usr/include/hdf5/serial LIBRARY_DIRS += /usr/lib/x86_64-linux-gnu/hdf5/serial |
and build it:
|
1 |
make && make distribute |
used video_stream_opencv to stream your video:
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 |
<!-- Example of run: roslaunch video_file.launch video_file:=/home/behnam/Desktop/dash_cam.mp4 --> <launch> <!-- launch video stream --> <!--remaping the topic --> <remap from="/videofile/image_raw" to="/image_raw"/> <include file="$(find video_stream_opencv)/launch/camera.launch" > <!-- node name and ros graph name --> <arg name="camera_name" value="videofile" /> <!-- full path to the video file --> <arg name="video_stream_provider" value="$(arg video_file)" /> <!-- throttling the querying of frames to --> <arg name="fps" value="30" /> <!-- setting frame_id --> <arg name="frame_id" value="videofile_frame" /> </include> </launch> |
download the trained model from here and put them in the model …
Car Detection Using Fast Region-based Convolutional Networks (R-CNN) in ROS with Caffe
To run this, you need to install Fast-RCNN and Autoware. Just in case you got error regarding hd5f when making Fast-RCNN, add the following lines to your Makefile.config
|
1 2 |
INCLUDE_DIRS += /usr/include/hdf5/serial LIBRARY_DIRS += /usr/lib/x86_64-linux-gnu/hdf5/serial |
Now run the following command to start:
|
1 2 3 4 |
source /opt/ros/kinetic/setup.bash source /home/<user_name>/Autoware/ros/devel/setup.sh export ROS_PACKAGE_PATH=/home/<user_name>/Autoware/:$ROS_PACKAGE_PATH roslaunch cv_tracker rcnn.launch |
if you got an error like :
|
1 |
Check failed: error == cudaSuccess (30 vs. 0) unknown error |
That means your graphics card is not ready or accessible, in my everytime I suspend my notebook I get that …
Car Detection Using Fast Region-based Convolutional Networks (R-CNN) in ROS with Caffe Read More »
Ackermann steering car robot model with simulation in Gazebo
Most of the wheeled robots in ROS use move_base to move the robot. move_base geometry model is based on differential drive which basically transforms a velocity command (twist message) into a command for rotating the left and the right wheels at a different speed which enable the car to turn into the right or left or goes straight. But cars …
Ackermann steering car robot model with simulation in Gazebo Read More »
Autonomous navigation of two wheels differential drive robot in Gazebo
Two wheels differential drive robot (with two caster wheels). List of installed sensors: • Velodyne VLP-16. • Velodyne HDL-32E. • Hokuyo Laser scanner. • IMU. • Microsoft Kinect/Asus Xtion Pro. • RGB Camera. You can manually control the robot with Joystick controller for mapping robot environment. Autonomous navigation is possible by setting goal pose.
