Path planning is an important topic in search and rescue for finding lost and injured persons in a short amount of time. For this we developed a new algorithm with maximizes the accumulated probability of target detection while
The main AOS repository with a list of related publications can be found here:
The code was written int Eclipse IDE with the integrated automated build system. Simply import the code into your workspace and start using it.
The configuration window offers a lot of parameters to specifiy. First of all, the path finding algorithms one would like to compare should be selected. Next, the user can choose from various different probability distributions to use for the simulations. Besides that, parameters like starting position can be choosed as well as parameters concerning the drone (speed and acceleration).
After all simulations have been computed, an overview is shown. For each algorithm, the found path defining the cell visiting order (discrete cells) is shown. Double clicking reveals another window as explained later. On the bottom the APT graph is shown for all compared algorithms.
This windows allows a direct comparison between the found path and the planned trajectory. The trajectory includes acceleration and deceleration segments and is used for calculating the total duration.
The code of our main algorithms can be found here:
- Grid: src/at/jku/cg/sar/pathfinder/misc/GridFinder.java
- Spiral: src/at/jku/cg/sar/pathfinder/misc/SpiralFinder.java
- Potential Field: src/at/jku/cg/sar/pathfinder/apf/AttractionApproachGradient.java
- Radial Gradient Accent: src/at/jku/cg/sar/pathfinder/misc/RadialChecker.java
- Continuous Gradient: src/at/jku/cg/sar/pathfinder/vacuum/VacuumCleanerGradient.java
Code concerning our trajectory planner can be found here:
- src/at/jku/cg/sar/trajectory/TrajectoryPlanner.java
- src/at/jku/cg/sar/trajectory/SimpleTrajectory.java