Method and a system for estimating the impact area of a military load launched from an aircraft
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
G06N-003/02
G06N-003/08
출원번호
US-0596416
(2008-04-09)
등록번호
US-8260734
(2012-09-04)
우선권정보
IT-TO2007A0272 (2007-04-18)
국제출원번호
PCT/IB2008/051341
(2008-04-09)
§371/§102 date
20091016
(20091016)
국제공개번호
WO2008/129435
(2008-10-30)
발명자
/ 주소
Tranchero, Bruno
Giura, Alberto
Converso, Stefania
출원인 / 주소
Alenia Aeronautica S.p.A.
대리인 / 주소
Merchant & Gould P.C.
인용정보
피인용 횟수 :
3인용 특허 :
2
초록▼
A system and a method are for the estimation of the impact area of a smart load that can be launched from an aircraft as a function of data or signals indicative of the aircraft flight conditions upon release of the load and of predetermined impact conditions on the target. The estimation of a polyg
A system and a method are for the estimation of the impact area of a smart load that can be launched from an aircraft as a function of data or signals indicative of the aircraft flight conditions upon release of the load and of predetermined impact conditions on the target. The estimation of a polygonal impact area defined by the coordinates of a central point and of a predetermined number of vertices is by corresponding neural networks.
대표청구항▼
1. A system for estimating the impact area of a smart load that can be launched from an aircraft, comprising a processing assembly including: means for acquiring first data or signals indicative of the aircraft flight conditions upon release of the load; andprocessing means including a plurality of
1. A system for estimating the impact area of a smart load that can be launched from an aircraft, comprising a processing assembly including: means for acquiring first data or signals indicative of the aircraft flight conditions upon release of the load; andprocessing means including a plurality of neural networks, for estimating the target impact point of the loadwherein said processing assembly further includes means for acquiring second data or signals indicative of predetermined conditions of impact on the target, andwherein the processing means are arranged for determining a polygonal impact area as a function of the first and second data or signals, said processing means including a plurality of first neural networks, arranged for calculating the coordinates of a central point of the impact area and of a predetermined number of vertices of the area, respectively. 2. A system according to claim 1, wherein the first neural networks are arranged for the calculation of the radii which extend from the central point of the impact area, which are adapted to define the predetermined number of vertices. 3. A system according to claim 1, wherein the processing means include a plurality of second neural networks which are arranged for the calculation of the time of flight to the vertices of the impact area. 4. A system according to claim 1, wherein the processing means for determining an impact area include a plurality of third neural networks which are arranged to estimate the existence of the impact area as a function of the first and second data or signals indicative of the aircraft flight conditions upon release of the load and of the predetermined conditions of impact on the target. 5. A system according to claim 4, wherein the first, second and third neural networks are backpropagation neural networks with from 3 to 5 layers and comprise transfer functions between the layers which are selected from the group comprising linear, sine, hyperbolic tangent, Gaussian, and complementary Gaussian transfer functions. 6. A system according to claim 5, wherein the first, second and third neural networks comprise 5-layer backpropagation neural networks with 3 hidden layers, having a hyperbolic tangent transfer function of the first hidden layer, a Gaussian transfer function of the second hidden layer, a complementary Gaussian transfer function of the third hidden layer, and a linear output transfer function, respectively. 7. A system according to claim 1, wherein the processing means are arranged for reducing the value of the radii of the impact area particularly for applying a first factor for the reduction of the front radii of the impact area, with respect to the direction of travel of the aircraft, and for applying a second factor for the reduction of the rear radii of the impact area, for calculating the coordinates of the predetermined number of vertices of a reduced impact area, and for calculating the time of flights of the load to the vertices of the reduced impact area. 8. A method of estimating the impact area of a smart load that can be launched from an aircraft, the method being adapted to be implemented by an electronic processing assembly comprising means for determining a polygonal impact area defined by the coordinates of a central point and of a predetermined number of vertices, as a function of first and second data or signals indicative of the aircraft flight conditions upon release of the load and of predetermined conditions of impact of the target, respectively, the means including a plurality of first neural networks arranged for calculating the coordinates of the central point of the impact area and of the predetermined number of vertices of the area, respectively,the method comprising:a learning stage in which the processing assembly receives first and second input training data or signals indicative of the aircraft flight conditions upon release of the load and of the predetermined conditions of impact on the target respectively, as well as associated output training data indicative of the coordinates of a central point and of a predetermined number of vertices of an impact area, and determines the transfer functions between the layers of the first neural networks as a function of the input and output training data, andan execution stage in which the processing assembly receives, as inputs, first and second measurement data or signals and calculates an impact-area estimate as a result of the application of the transfer functions. 9. A method according to claim 8, wherein the first neural networks are arranged for calculating the radii which extend from the central point of the impact area, which are adapted to define the predetermined number of vertices. 10. A method according to claim 8, wherein the means include a plurality of second neural networks arranged for calculating the time of flight to the vertices of the impact area, respectively, the method comprising:a learning stage in which the processing assembly receives first input training data or signals indicative of the aircraft flight conditions upon release of the load and second input training data or signals indicative of the predetermined conditions of impact on the target, as well as associated output training data indicative of the time of flight to the vertices of the impact area, and determines the transfer functions between the layers of the second neural networks as a function of the input and output training data, andan execution stage in which the processing assembly receives, as inputs, first and second measurement data or signals and calculates an estimate of time of flight to the vertices of the impact area as a result of the application of the transfer functions. 11. A method according to claim 8, wherein the means include a plurality of third neural networks arranged for estimating the existence of the impact area, respectively, the method comprising:a learning stage in which the processing assembly receives first input training data or signals indicative of the aircraft flight conditions upon release of the load and second input training data or signals indicative of the predetermined conditions of impact on the target, as well as associated output training data indicative of the existence of an impact area, and determines the transfer functions between the layers of the third neural networks as a function of the input and output training data, andan execution stage in which the processing assembly receives, as inputs, first and second measurement data or signals and determines the presence or absence of an impact area as a result of the application of the transfer functions. 12. A method according to claim 8, wherein the first, second and third neural networks are backpropagation neural networks with from 3 to 5 layers and comprise transfer functions between the layers which are selected from the group comprising linear, sine, hyperbolic tangent, Gaussian, and complementary Gaussian transfer functions. 13. A method according to claim 12, wherein the first, second and third neural networks comprise 5-layer backpropagation neural networks with 3 hidden layers having a hyperbolic tangent transfer function of the first hidden layer, a Gaussian transfer function of the second hidden layer, a complementary Gaussian transfer function of the third hidden layer, and a linear output transfer function, respectively. 14. A method according to claim 1, comprising reducing of the value of the radii of the impact area, calculating of the coordinates of the predetermined number of vertices of a reduced impact area, and calculating of the time of flights of the load to the vertices of the reduced impact area, wherein the reduction of the value of the radii of the impact area comprises applying of a first reduction factor to the front radii of the impact area, with respect to the direction of travel of the aircraft, and applying of a second reduction factor to the rear radii of the impact area. 15. A processing program or set of programs adapted to be executed by a processing system and which comprise/s one or more code modules for the implementation of the neural networks of a system for estimating the impact area of a smart load according to claim 1.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (2)
Krogmann, Uwe, Method of training a neural network for the guidance of a missile to a target.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.