Using various sensing modalities (distance, AOA, TDOA, hybrid);
Perform significantly better than known alternatives, more robust and guaranteeing better precision;
Some for precise localization of targets on (near-) spherical surfaces, with verification analysis using real-life trial data.
Identification of sensing/control architecture design principles
To ensure preservation of the shape of a formation;
Decentralized implementation/use is an issue.
Set of formation control algorithms
To maintain the shape of a three vehicle formation in the presence of wind, constraints on airspeed, turning rate, etc;
Desired shape is important in maximizing accuracy of cooperative localization with sensors on these vehicles;
Obstacle avoidance.
Robust self-localization algorithms for three (or more) UAV formations
A particular one based on inter-agent distance and bearing from two land marks information;
Self-localization and tracking algorithms for UAV formations in loss of GPS measurement.
Identification of sensor-target geometric principles in cooperative localization
For measuring/detecting/avoiding (near-) collinearity problems;
For sensor positioning to optimize quality of target localization.
Progress Update for SWARM
The 3-year NICTA-DSTO project SWARM was completed in October 2008. All scheduled milestones have been achieved. End-of-project reports have been provided to both NICTA and DSTO. The list of publications of the project team members related to the SWARM project can be reached from the main page.
Key Research Outcomes to Date of SWARM-II
Outcome
Details
Close-target reconnaissance using distance and bearing-only measurements
A control algorithm for close-target reconnaissance was developed using adaptive control techniques;
The algorithm was designed to use either distance-only or bearing-only measurements;
An adaptive target localization algorithm was developed for use with bearing-only measurements;
The algorithm has been extended to account for mobile targets.
Discrete probability density maps
Developing algorithms for the evolution of discrete probability density maps.
Target de-interleaving via time-difference-of-arrival
We are excited about a paper published with DSTO in Applied Physics Letters in 2009;
A patent on this work is currently being applied for;
DSTO are constructing experimental gear for ground trials and then aim to instrument UAVs for flight trials, possibly in a US competition.
Cooperative target localization
A general approach was discovered for calculating the bias in localization problems;
Outperforms best known methods for bias elimination in bearing-only localization.
Progress Update for SWARM-II
DSTO is building a prototype for de-interleaving signals from multiple emitters, and the theoretical foundation of the algorithm is originated from the SWARM project (the predecessor project 2005-2008). A NICTA-DSTO joint paper has appeared in Applied Physics Letters.
A few formation flying algorithms, first developed at NICTA for the DSTO’s flight simulator,are scheduled to be tested by DSTO in upcoming trials.
NICTA-developed localization algorithms have also been tested on trial data collected by DSTO in the USA.
