Novel alghorithm for autonomous indoor UAV orientation and position estimation based on a 2D lidar, IMU and a 1D lidar.

This algorithm has been created for my Bachelor graduation internship at TNO. The assignment was to develop a small-as-possible drone that can estimate its own position and orientation in an indoor envirpnment.

Main contributions:

• Efficient line extraction alghorithm (split-and-merge) based on 2D LIDAR
• Filters for walls selection (line filtering) and wall matching
• Accurate yaw estimation
• Frist draft of position estimation
• Kalman filter that is able to 1) estimate the IMU bias and 2) to smoothen the LIDAR estimate

(gif: the left image is the previous (t-1) LIDAR reading and the right image the current (t) reading. The yellow lines are basically the LIDAR rays, so the the length of the yellow lines are the distance readings. Notice the red line being tracked between every LIDAR reading in real-time; meaning that we are able to estimate our rotation/yaw based on the changing orientation towards this (red) wall)

Setup

This algorithm can either be tested with a real LIDAR and Pixhawk 4 or simply be tested with pre-recorded data avaible in this repo.

Go to the algorithm directory from this repo (all contributions are stored here):

cd ./poseEstimationBasedOnLidar

A. Test with pre-recorded data

Open the main code file (poseEstimationBasedOnLIDAR.py) in your favorite text editor. From terminal:

xdg-open poseEstimationBasedOnLIDAR.py

Then make sure URGmocker and pixhawk4 (lines 60 and 61) are constructed with the argument "READ_FROM_FILE", like this:

self.mocker = URGMocker(READ_FROM_FILE)
self.pixhawk4 = pixhawk(READ_FROM_FILE)

Then open mockLIDAR.py and comment line 12, like so:

#from breezylidar import URG04LX

Then run the program. From terminal:

./poseEstimationBasedOnLidar

B. Testing with real hardware (Hakuyo URG04LX and pixhawk)

0. Install BreezyLIDAR

https://github.com/simondlevy/BreezyLidar (Linux only)

1. Install ROS

http://wiki.ros.org/kinetic/Installation/Ubuntu I did those steps on ubuntu 16.04:

sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main" > /etc/apt/sources.list.d/ros-latest.list'
sudo apt-key adv --keyserver hkp://ha.pool.sks-keyservers.net:80 --recv-key 421C365BD9FF1F717815A3895523BAEEB01FA116
sudo apt-get update
sudo apt-get install ros-kinetic-desktop-full
sudo rosdep init
rosdep update
echo "source /opt/ros/kinetic/setup.bash" >> ~/.bashrc
source ~/.bashrc
sudo apt install python-rosinstall python-rosinstall-generator python-wstool build-essential

2. Install mavROS

Simplest install is for Ubuntu, https://dev.px4.io/en/ros/mavros_installation.html

sudo apt-get install ros-kinetic-mavros ros-kinetic-mavros-extras
wget https://raw.githubusercontent.com/mavlink/mavros/master/mavros/scripts/install_geographiclib_datasets.sh
./install_geographiclib_datasets.sh

3. Configure static usb ports

First configure static usb ports for the hardware devices: lidar to /dev/URG04LX and pixhawk4 to /dev/pixhawk4 , use these IDS:

static usb reference for pixhawp4:
    idVendor: 26ac
    idProduct: 0032
static usb reference for URG04LX:
    idVendor: 15d1
    idProduct: 0000

Alternatively you could edit the code to just use the dynamic assigned usb names (like /deb/USBtty0) to the pixhawk or LIDAR

4. Let's run it

run px4.launch:

roslaunch mavros px4.launch fcu_url:=/dev/pixhawk4:57600  gcs_url:=udp://@localhost:14550

check if u see imu data incoming:

rostopic echo /mavros/imu/data

then run:

./poseEstimationBasedOnLIDAR.py

4. Configure algorithm to listen to usb devices

OPTIONAL: Only if you tested before with pre-configured dataset, then you have to set pixhawk and URGmocker constructor arguments back to "READ_FROM_SERIAL" in poseEstimationBasedOnLIDAR.py:

self.mocker = URGMocker(READ_FROM_SERIAL)
self.pixhawk4 = pixhawk(READ_FROM_SERIAL)

And uncomment line 12 from mockLIDAR.py:

from breezylidar import URG04LX

NOTE, if this programs returns breezylidar connect error, then just retry 2 times