A method of guiding a pursuer to a target is provided, and is of particular use when the possible target location is described by non-Gaussian statistics. Importantly, the method takes into account the fact that different potential target tracks in the future have significantly different times to go
A method of guiding a pursuer to a target is provided, and is of particular use when the possible target location is described by non-Gaussian statistics. Importantly, the method takes into account the fact that different potential target tracks in the future have significantly different times to go. That can give rise to emergent behavior, in which the guidance method covers several possible outcomes at the same time in an optimal way. An example embodiment of the method combines Particle Filter ideas with Swarm Optimization techniques to form a method for generating guidance commands for systems with non-Gaussian statistics. That example method is then applied to a dynamic mission planning example, to guide an airborne pursuer to a ground target travelling on a network of roads where the pursuer has no-go areas, to avoid collateral damage.
대표청구항▼
1. A method of guiding a pursuer to a target, the method comprising: (a) providing a plurality of possible target trajectories, weighted to represent a probability distribution of the target trajectories;(b) providing a plurality of candidate guidance controls, parameterised by a candidate guidance
1. A method of guiding a pursuer to a target, the method comprising: (a) providing a plurality of possible target trajectories, weighted to represent a probability distribution of the target trajectories;(b) providing a plurality of candidate guidance controls, parameterised by a candidate guidance parameter;(c) for each of a plurality of values of the candidate guidance parameter: a. determining a projected pursuer trajectory resulting from the candidate guidance control corresponding to the candidate guidance parameter value;b. using the projected pursuer trajectory so far developed, determining the next occurrence of a termination condition for the pursuer trajectory relative to any of the target trajectories for which a termination condition has not yet occurred;c. adjusting the projected pursuer trajectory, subsequent to that termination-condition occurrence, to account for that target trajectory;d. repeating steps b and c until the termination conditions corresponding to all of the plurality of possible target trajectories have been accounted for; ande. recording a cost for the candidate guidance control based on the projected pursuer trajectory and the miss distances at the termination-condition occurrence for each target trajectory;(d) generating a revised plurality of values of the candidate guidance parameter in view of the recorded costs of the candidate guidance controls;(e) repeating steps (c) and (d) a plurality of times and identifying an optimal guidance control based on the recorded costs;(f) applying the optimal guidance control to the pursuer to alter its actual trajectory;(g) obtaining updated possible target trajectories; and(h) repeating steps (b) to (g) a plurality of times. 2. A method as claimed in claim 1, in which the termination condition relative to a target trajectory is the point of nearest approach of the projected pursuer trajectory to the target trajectory. 3. A method as claimed in claim 2 wherein, for said plurality of possible target trajectories provided in step (a), the time of flight to a corresponding point of nearest approach for the pursuer trajectory relative to each of the target trajectories shows a spread of the order of 30% or more with regard to the mean time of flight for all the target trajectories. 4. A method as claimed in claim 1, in which the candidate guidance parameter is a costate parameter. 5. A method as claimed in claim 1, in which the target trajectories are described by non-Gaussian statistics. 6. A method as claimed in claim 5, in which the target trajectories are possible routes of a target vehicle. 7. A method as claimed in claim 1, in which the revised plurality of values of the candidate guidance parameter are generated using an optimisation algorithm. 8. A method as claimed in claim 1, in which the optimisation algorithm is a swarm algorithm. 9. A method as claimed in claim 8, in which the swarm algorithm is a master-slave swarm algorithm. 10. A method as claimed in claim 1, comprises uncurling a projected trajectory by damping a high-frequency oscillation in the projected trajectory. 11. A method as claimed in claim 1, in which said plurality of possible target trajectories provided in step (a) comprise a number of possible discrete trajectories. 12. A method according to claim 11, in which said possible discrete trajectories comprise road routes. 13. A method as claimed in claim 1, further comprising providing one or more pursuer state constraints such that each cost recorded in step (e) for the candidate guidance control is subject to said pursuer state constraint(s). 14. A system for guiding a pursuer to a target, the system comprising: a processing apparatus configured to obtain a plurality of possible target trajectories, weighted to represent a probability distribution of the target trajectories; and a plurality of candidate guidance controls, parameterised by a candidate guidance parameter;the processing apparatus being configured to perform the steps of:(a) for each of a plurality of values of the candidate guidance parameter, determining a projected pursuer trajectory resulting from the candidate guidance control corresponding to the candidate guidance parameter value;using the projected pursuer trajectory so far developed, determining the next occurrence of a termination condition for the pursuer trajectory relative to any of the target trajectories for which a termination condition has not yet occurred;adjusting the projected pursuer trajectory, subsequent to that termination-condition occurrence, to account for that target trajectory;repeating steps b and c until the termination conditions corresponding to all of the plurality of possible target trajectories have been accounted for; andrecording a cost for the candidate guidance control based on the projected pursuer trajectory and the miss distances at the termination-condition occurrence for each target trajectory;(b) generating a revised plurality of values of the candidate guidance parameter in view of the recorded costs of the candidate guidance controls;(c) repeating steps (c) and (d) a plurality of times and identifying an optimal guidance control based on the recorded costs;(d) applying the optimal guidance control to the pursuer to alter its actual trajectory;(e) obtaining updated possible target trajectories; and(f) repeating steps (b) to (g) a plurality of times. 15. The system of claim 14 wherein the pursuer is a vehicle comprising the processing apparatus. 16. The system of claim 15 wherein the vehicle comprises a munition. 17. The system of claim 14 further including a ground based radar station comprising the processing apparatus.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (21)
Yueh William R. (Claremont CA), Augmented proportional navigation in second order predictive scheme.
van Tooren, Joost; Heni, Martin; Knoll, Alexander; Beck, Johannes, Collision and conflict avoidance system for autonomous unmanned air vehicles (UAVs).
Yueh William R. (Claremont CA), Fourth order predictive, augmented proportional navigation system terminal guidance design with missile/target decouplin.
Yost David J. (Woodbine MD) Arrow Arthur (Beltsville MD) Konigsberg Robert L. (Baltimore MD), High angle-of-attack missile control system for aerodynamically controlled missiles.
Boelitz, Frederick Wall; Singh, Leena; Gibson, Christopher Michael; Smith, Jr., Stephen Randolph; Miotto, Piero; Martin, Andrew Allen, Load relief system for a launch vehicle.
Aragones, James Kenneth; Kiaer, Carol Lynn; Maruscik, Ronald George; Taware, Avinash Vinayak; Mattheyses, Robert Marcel, Methods and apparatus for reducing hyperplanes in a control space.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.