The Research On Beam And Power Managemengt Of Integrated Sensing And Communication

Integrated sensing and communication(ISAC)technology is one of the key technologies in B5G/6G,which provides users with two services,sensing and communication.The mobile network will a converged network of sensing and communication in the foreseeing future.ISAC utilizes mmWave and mMIMO to get higher transmission rates and higher measurement resolution.Therefore,addressing the performance problem of emerging applications such as vehicle-ground collaboration.But due to the high mobility the performance overhead is heavily dependent on the alignment of the beam.Maintaining real-time beam alignment in High mobility scenarios will be a huge challenge.Secondly,for communication narrower beams is better,On the contrary,a wider beam is better.It will be a great challenge to reconcile the user’s communication and perception needs in a highly mobility scenario by adjusting the beam width.Finally,because the high mobility of the users will lead to frequent channel state changes and consequently frequent performance fluctuations.Allocating resources for better network performance is a great challenge.In literature,research on beam management in ISAC systems is mainly focused on beam tracking under the assumption of point targets,while beam management under the assumption of extended targets,which is closer to the actual scenario,has rarely been discussed.Research on resource allocation for ISAC systems has focused on unifying communication and sensing metrics,and developing metrics that can be used to measure both communication and sensing performance to ensure optimal system performance.However,there has been a lack of discussion on the fairness of performance achieved across multiple users.It is also important to maintain user fairness.In this paper,the beam management and resource allocation aspects of the ISAC system are therefore investigated in the following ways.(1)In this paper,a deep learning-based beam management scheme for ISAC systems under the extended target assumption is first proposed for V2I scenarios to improve the performance under the extended target assumption.Specifically,a Gated Recurrent Unit-based beam tracking method is proposed to achieve efficient and accurate beam alignment and obtain a utility function closer to that of the beam training method;then a Deep Neural Network-based beamwidth control method is proposed to achieve efficient beamwidth control and obtain a utility function closer to that of the exhaustive search method.This ensures better performance.(2)In this paper,we propose a fairness-guaranteed time and power resource allocation scheme in the ISAC system,which improves the users fairness.Specifically,this paper first derives sensing fairness,communication fairness metrics among multiple users and weighted sum of the sensing-communication MMSE based on the vehicle’s speed/range/pitch azimuth Cramer Lower Bound and distortion minimum mean square error metrics as communication metrics.An optimisation model is then developed with the objective function of minimising the weighted sum of the three,and the independent variables of allocating power among multiple users and allocating the proportion of time in the sense frame phase to sense subframes for sense users.This paper significantly improves user fairness at the expense of only a small amount of performance,which is The weighted sum of the sensing-communication MMSE.An improved alternating iterative power and time resource allocation algorithm is then proposed to improve the speed of model solving.Finally,the paper provides a brief summary of all the research and an outlook on the subsequent research in sensorimotor fusion systems.