Home | Members | Research | Facilities | Publications | Software | News | Contact


 

   

Formation Flight and Swarms

 
Vehicle Formation

Many leader-follower methods have been developed for vehicle formation flying.  Some of the earlier studies used simplified point mass models in their developments, but some of the more recent studies have used vehicle models that do not allow side-slip, which mimics an air vehicle's inability to naturally have side slip motion.   

Rendezvous can also entail formation flying or interception problems where the origin is effectively moving. Interception of incoming ballistic missiles is a rendezvous problem where the origin becomes a moving target and one of the agents is non-cooperating. Formation flying is a type of rendezvous problem where multiple agents must coordinate position and velocity. The docking of two spacecraft is a rendezvous problem that involves the two spacecraft matching both position and velocity with the proper orientation. Air-to-air refueling is another rendezvous problem. Additional applications arise in submersibles where robotic vehicles must converge upon a set location, either moving or stationary.

Leader Follower UAV formation pairing

fig 10. Illustration of leader follower UAV formation pairing

Vehicle Swarms
Researchers have been developing cooperative agent-based methods to generate solutions to many types of problems.  These methods are characterized by a team of agents that sample and share information to generate feasible solutions to complex problems.  Agent can mean a physical entity, like a robot, or a non-physical entity, like a trial solution to a design problem.  Some example platforms and applications include robotic vehicles that cooperate to locate buried land mines using exhaustive searches and software agents that perform information search and retrieval on the world-wide web.  One primary focus has been on cooperative search and localization algorithms for robot collectives. 

Our research has generated robust cooperative control methods for a robotic vehicle team that is tasked with localizing a time-invariant, stationary source that emits a measurable scalar field.  We assume that each robot collects information about the time-invariant field through a (set of) sensor(s) and this information, together with a robot's position, is shared among the robots through a communication network.  This measuring and sharing of information means that the robots represent a distributed sensor network.  Each robot then uses the information it gathers and receives to autonomously determine its own position update which will take it closer to the unknown source location.  Consequently, the robot controls are decentralized feedback controls.  The development of the control law for each robot is inspired by classic function-minimization theory.

Figure 12 shows an illustration of cooperative localization among nine vehicles.  The vehicles are implementing a strategy of collecting and sharing information, and determining their own control action.  The vehicles movement is from X to O, and one see that localization of the unknown source is accomplished.  Figure 13 shows a source localization scenario for a multimodal surface.  Again a team of vehicles are tasked with localizing a source, as given by the minimum of the surface. Figure 14 shows results for the case of global communication, wherein all vehicles are able to communicate.  Using this communication strategy, the global minimum is found by the vehicles. Figure 15 shows results for the case of local communication, wherein the vehicles can only communicate to a set of nearest neighbors.  Using this communication strategy, the local minimums are found by the vehicles.

Performance of full state leader follower

Figure 11: A qualitative comparison of full state leader-follower control (left) and a velocity-free leader follower control (right).

 

 

Source Localization


Figure 12: An illustration of cooperative localization. 

Source Localization1


Figure 13: Multimodal surface for source localization example.

Source Localization2
Figure 14: Cooperative localization on a multimodal surface with global communication.

 

Source Localization3

Figure 15: Cooperative localization on a multimodal surface with local communication.

PIs: John E. Hurtado, Raktim Bhattacharya, Others
 

Publications :

  1. A New Perspective on Decentralized Cooperative-Control Design for Vehicle Formations - L.A. Weitz, J.E. Hurtado, A.J. Sinclair, AIAA 2007 GNC Conference, August 2007. 

  2. Convergence of Newton’s Method via Lyapunov Analysis  - J.E. Hurtado, R.D. Robinett III, Journal of Guidance, Control, & Dynamics, Vol. 28, No. 2, 2005, pp. 363-364.

  3. Decentralized Control for a Swarm of Vehicles Performing Source Localization - J.E. Hurtado, R.D. Robinett III, C.R. Dohrmann, S.Y. Goldsmith, Journal of Intelligent & Robotic Systems: Theory and Applications, Vol. 41, 2004, pp.1-18.

  4. Stability and Control of Collective Systems - R.D. Robinett III, J.E. Hurtado, Journal of Intelligent & Robotic Systems: Theory and Applications, Vol. 39, 2004, pp. 43-55.

  5. Cooperative Control of Vehicle Swarms for Acoustic Target Localization by Energy Flows - J.L. Dohner, G.R. Eisler, B.J. Driessen, J.E. Hurtado, Journal of Dynamic Systems, Measurement, and Control, Vol. 126, Issue 4, 2004, pp. 891-895

  6. Miniature Mobile Robots for Plume Tracking and Source Localization - R.H. Byrne, D.R. Adkins, S.E. Eskridge, J.J. Harrington, E.J. Heller, J.E. Hurtado, J.E., Journal of Micromechatronics, Vol. 1, No. 3, 2002, pp. 253-261.



 


© Aerospace Engineering, Texas A&M University