Research On Cognitive Collaboration Technology Of Distributed Radars

With the single radar's increasing prominent performance bottleneck and the dilemma in dealing with future new threats,the radar sensing paradigm shift is taking place from the current monostatic radar to the geographically distributed radars collaboration.Comparing with the single monostatic radar system,distributed radars collaboration can gather target scattering from multiple dimensions such as space,frequency,and polarization,thus it can offer the potential to extend the capabilities and performance of radar systems.However,the collaborative approach of multi-radars and the choice of the radar's parameters directly affect the performance of collaboration.Hence,it is necessary to adaptly schedule the radar in collaboration in responding to the varying scenarios,that is,to achieve intelligent collaboration.The feasible idea to realize the intelligence is to imitate the cognitive-decision cycle,that is,determining the behavior of radar via the feedback(cognitive knowledge)via interacting with the scenario.Inspired by this,how to design interaction and cognition methods and decision-making schemes for specific collaboration purposes are challenging,and it is also a technical bottleneck to be broken through urgently in the future.Under this background,two specific aspects of intelligent collaboration: cognitive coherent combination and cognitive resource allocation are discussed in this dissertation.The main contributions of the dissertation are as follows:1.The theory and the boundary conditions of cognitive coherently combining distributed multiple radars.Coherently combining distributed multiple radars aims to obtain enhanced received signal strength with the idea of “smaller accumulation”.For this collaboration purpose,the essence of coherently combing is revealed firstly,that is,coherently combining is equivalent to adjusting each radar's transmit/receive time and phase to calibrate the incoherence introduced by radar distributed arrangement.On this basis,the cognitive framework for perceiving the coherent calibrated values,i.e.,the coherent parameters(CPs),via interacting with the target is proposed.However,the CPs perceived always accompanied by uncertainty.Hence,the theoretically achievable performance gain upper bound after coherently combining multiple radars with uncertainties in CPs is discussed in depth.Besides,the influence of multi-radar geometric arrangement on the performance of coherently combining is also discussed.And two constraints of radar geometrical arrangement,i.e.,the signal coherence constraint and the CPs substitutable constraint,are derived under the premise of obtaining considerable or expected coherently combining benefits.2.The implementation methods and processing flow of cognitive coherently combining distributed multiple radars.Under the established basic theoretical prototype of coherently combining multiple radars,the new problems faced by the more practical cognitive coherently combining scenarios are discussed,including: the biased or noise-sensitive problem of transmit CPs estimation when transmitting non-ideal“interacting” orthogonal waveforms,the unreachable problem of transmit CPs when transmitting the identical coherently combining signal and the hysteresis problem in perceiving transmit CPs when the “interacting” target is moving.And the responding solutions are proposed independently: the peak picking-oriented solution based on designing the orthogonal polyphase coded waveforms with dexterous strong suppression interval and the universal solution based on “clean” separation echo reconstruction;the indirect perceiving method via receive CPs measurement conversion,and the CPs prediction scheme for coping with the hysteresis.The above researches aim to give the solutions of perceiving CPs for achieving coherent combining under the more general scenarios other than the theoretical prototype scenario and maintain this coherently combining state.Furthermore,a general staged processing flow for coherently combining multiple distributed radars is proposed by integrating all the key technologies discussed above,whose validity is verified by a dual-radar coherently combining example implemented under a ballistic target tracking scenario;3.Cognitive resource allocation in distributed multiple radars collaborative tracking.The purpose of resource allocation is to predictively control the generation of radar measurements to better perform collaborative tasks.Focusing on this problem,the factors affecting the target localization accuracy are discussed firstly from the perspective of the likelihood function and the Fisher information matrix(FIM)qualitatively and quantitatively.And the collaborative positioning optimality conditions are derived to demonstrate the necessity of resource allocation;On this basis,a general cognitive resource allocation framework for collaborative tracking with multiple distributed radars is designed.The distinctive feature is developing the predicted conditional Cramer-Rao bound(PC-CRLB)as an evaluation metric by introducing the concept of virtual measurement.As the PC-CRLB can characterize the future response of candidate decisions conditioned on the uncertain cognitive knowledge.The allocation strategy making is robust to cognitive knowledge.Besides,to verify the validity of the proposed allocation framework,the proposed allocation framework is applied to a dwell time allocation problem in different contexts,and show that such a problem can be solved via cone programming.