crazyflie_human

ROS and crazyflie-integrated prediction of human motion.

Dependencies

This repository depends on the pedestrian_prediction repository which may be found here and the crazyflie_clean package which may be found here.

Usage

This repository needs to be placed inside of a ROS workspace. Clone it inside of the src/ folder in your workspace:

cd catkin_ws/src/
git clone https://github.com/abajcsy/crazyflie_human

Simulated Human Motion

To run predictions with simulated human data, type:

roslaunch crazyflie_human simulated_demo.launch

Changing the Simulated Human Model

There are many ways to simulate human pedestrian data. This repository supports:

• linear_human.py -- Human motion is simply a straight line between goals. Ignores obstacles.
• potential_field_human.py -- Human motion follows attractive-repuslive forces towards goals and away from obstacles.

Once you know which simulation you want to use, open simulated_human_launcher.launch. Change the simulation node to point to the appropriate python file. Now just rerun the simulated demo!

Real (Optitrack) Human Motion

To run predictions with real human data, type:

roslaunch crazyflie_human real_demo.launch

Changing the Number of Humans

Open up either simulated_demo.launch or real_demo.launch. To add another human into the launch file, simply add a new namespace for this human, following the convention:

<arg name="humanN_namespace" default="humanN" />

then add a new human ID:

<arg name="humanN" default="N" />

then add a new human prediction node:

<group ns="$(arg humanN_namespace)">
	<include file="$(find crazyflie_human)/launch/real_human_launcher.launch">
		<arg name="human_number" value="$(arg humanN)" />
		<arg name="beta" value="$(arg prediction_model)" />
	</include>
</group>

Changing the Human Start and Goals

Open /config/pedestrian_pred.yaml. For each human (numbered 1-N) make sure that they each have specified starts and goals:

humanN_real_start: [1.0,-1.0]
humanN_real_goals: [[1.0, 2.0], [2.0, 1.0]]

These are used for predicting future states of the human, and are used by the human simulator to move the simulated human through the environment.

Visualization

To see the visualizations of the human, the predictions, and the modeled goals, run RVIZ:

roslaunch crazyflie_human rviz.launch