Study On Resource Allocation Scheme For Cognitive Radar

Cognitive radar, as the next generation intelligent radar, is an important developing direction of the radar in the future. The main character of cognition is a feedback from the receiver to the transmitter, which makes it possible for the radar system to work on a cycle to cycle basis. The prior information about the environment and the target fed back from the tracker enable the radar system to adapt both the processing manner of the receiver, as well as the working parameter of the transmitter. In practice, the transmit resource of the radar system is limited in some specified application, such as transmit power and signal bandwidth. Therefore, the optimization of limited radar resource, using the idea of the cognition, is of critical importance. Under this background, this thesis mainly discusses how to adaptively design the detection threshold of the receiver and the working parameters of the transmitter according to the feedback information, in order to achieve better detection and tracking performance under the constraints of limited resources. According to the working platform and the number of the targets, this thesis divides the tracking problems into the following categories: single-radar single-target tracking, single-radar multi-target tracking, multi-radar single-target tracking and multi-radar multi-target tracking. With the fully understanding of relevant research both at home and abroad, this thesis conducts an in-depth research on the resource allocation scheme of cognitive radar. The main works done in this thesis can be listed as follows:1. This thesis studies a single-radar single-target cognitive tracking algorithm in dense clutter environment. With the aim of enhancing target detection and tracking performance, this thesis mainly discusses how to set the detection threshold in the validation gate with the use of the obtained target information.2. This thesis studies two cognitive multitarget tracking algorithms under both ideal and non-ideal detection condition for monostatic colocated MIMO radar. Firstly, this thesis establishes target motion, target detection and observation model. Based on which, the tracking BCRLB of the multiple targets are derived. Then, the mathematical model of resource allocation scheme is built, in order to minimize the worst case tracking BCRLB of the multiple targets. For the ideal detection case, this thesis considers a joint power and beam allocation scheme, and proves that the resulting optimization problem is equivalent to solving multiple convex problems. For the un-ideal detection case, this thesis develops a method to design the detection threshold of the receiver, as well as the transmit power of the multiple beams according to the feedback information.3. This thesis studies several resource allocation algorithms for case of multi-radar single-target tracking, in view of different application background.(1) Starts from different signal model, this thesis develops a power allocation algorithm for UCW radar network, and proves that power allocation is a convex optimization problem.(2) Aiming at the problem that the real time processing capacity of the fusion center is limited, this thesis develops a joint power and bandwidth allocation scheme.(3) This thesis develops a power allocation strategy for single target tracking in asynchronous MRS.(4) This thesis treats target RCS(Radar Cross Section) as a random parameter, and proposes a NCCP(Nonlinear Chance Constraint Programming) based robust power allocation algorithm.(5) In view of the 3D object, this thesis proposes a suboptimal tracking algorithm for distributed radar network. Based on which, a power allocation algorithm is developed for 3D object tracking.4. This thesis studies a joint cluster and power allocation algorithm for the case of multi-radar multi-target tracking. The general steps can be described as: select an optimal subset of radars with predetermined size to cluster around each target and implement the power allocation strategy among those selected radars. In this case, the MRS can adaptively adjust the transmit parameter and determine which measurement will be used, and thus can achieve better performance.