Systematization of Knowledge: Spectrum Sharing Between Radar and Communications

To help meet the growing demand for wireless data capacity, the United States Federal Communications Commission (FCC) and National Telecommunications and Information Administration (NTIA) define and regulate a spectrum-sharing approach in several radio frequency (RF) bands, with sharing in additional bands expected in the near future. This paper examines the case in which some of the users in the shared bands are government radars, as is the case in the 3.5 GHz Citizens Broadband Radio Service (CBRS) band. This paper surveys topics related to dynamic spectrum sharing (DSS) of RF bands between government radars and communication systems, including spectrum policy, dynamic spectrum access (DSA), dynamic frequency selection (DFS) and frequency assignment, Primary User (PU) protection and privacy, radar-communications interference mitigation, and joint radar-communications system design. The purpose of this paper is to provide readers with a summary of current knowledge of these aforementioned concepts while also providing references so that topics of interest may be studied deeper. The structure of this paper is broken into several topics, where within each topic table(s) provide a survey of the relevant topic. The contribution made includes over 200 references. Tables include summaries of approaches described in multiple references, as well as descriptions and summaries of individual references used throughout the survey. A taxonomy of approaches applicable to spectrum sharing between radar and communication systems is provided, as are conclusions and recommendations. Findings include existing technology, systems, and regulations utilizing the CBRS in its current structure. Therefore, all of the advancements will most likely be derived using a combination, derivative, or adapted methodology defined in the different sections throughout this survey.


I. INTRODUCTION
This paper provides a systematization of knowledge regarding Dynamic Spectrum Sharing (DSS) between federal radars and wireless communications systems within the United States.Although specific frequency bands, coexisting systems, and regulations considered may be unique to the US, the paper surveys approaches and technologies that are potentially applicable in any country.We survey viable techniques, provide a qualitative assessment of shared spectrum paradigms, and consider underlying technologies and protocols.Further, we recommend technologies and The associate editor coordinating the review of this manuscript and approving it for publication was Yafei Hou .
approaches that have promise given policy, regulation, and security issues, e.g., operational security.The topic is timely due to: (1) recently-initiated sharing of federal radar frequencies with non-federal wireless systems in the 3.5 GHz Citizens Broadband Radio Spectrum (CBRS) band; (2) planned expansion of spectrum sharing to lower frequencies in the 1.3 GHz and 3.1-3.55GHz bands that are also used by federal radars including airborne, ground-based and Naval shipboard radars; and (3) potential future sharing in other bands [1], [2], [3], [4].Topics and approaches covered in the paper are shown in Figure 1.DSS is a subsection of various types of frequency sharing.DSS can be implemented with radar systems, communication systems, and joint radar and communication systems.Non-cooperative systems are those where two or more communication systems coexist without knowledge of each other's architecture or operation and must sense the operations of the other to prevent interference.Cooperative models are when two or more communication systems work together to inform each other of their operation to prevent interference.Signal processing is a widely applicable topic extending beyond radar and communication systems.As shown in the figure, signal processing overlaps and extends beyond the boundaries of these topics.Cognitive Radio (CR) and cognitive communications are essential components of DSS and also function as distinct subsections within radar and communication systems, each with applications both within and outside the realm of DSS.Note that some approaches such as sensing and opportunistic sharing are intentionally omitted from the figure due to their extensive overlap with multiple approaches, rendering their representation impractical and potentially confusing for visual presentation.
Table 1 is given to facilitate understanding of the key terms and concepts in the field of cognitive radio, spectrum sharing, and communications.Table 1 is a comprehensive table of acronyms and definitions used throughout the paper.This table includes relevant governing organizations, processes, functions, and standardization efforts.
We present a taxonomy of DSS approaches and summarize research literature within several broad categories related to DSS between radar and communications.We summarize research in each category.For categories that include a large number of publications, we provide tables that compare and contrast specific sub-categories, common approaches, and specific solutions proposed in individual publications.We identify and recommend technologies that have potential for improving DSS between federal radar and both federal and non-federal wireless communications systems, and summarize policy, regulatory, and security issues that will influence the selection and development of these technologies.
Section I, this section, provides an overview of the topics, key survey papers to be discussed, and a brief introduction to the items covered in the remaining sections.
Section II gives examples of bands used by federal radars that are already shared or are candidates for sharing with communication systems.
Sections III, IV, and VI provide a progressively detailed exploration of DSS approaches starting with surveys, and ending with specific studies and approaches.
Section III overviews literature surveys related to DSS and provides a taxonomy of approaches.
Section IV discusses examples of more specific approaches to enable DSS that are applicable but not exclusive to spectrum sharing by or with radars.
Section V summarizes selected policy documents related to radar-communications spectrum sharing.
Section VI describes spectrum occupancy studies and DSS approaches.Section VI-A discusses centralized approaches to spectrum sharing.Section VI-B describes opportunistic sharing.Section VI-C describes approaches to spectrum sensing.
Section VII summarizes signal processing techniques for radars that enhance spectrum sharing with other systems.It also provides an overview of the concept of cognitive radar, which provides flexibility that can enhance sharing with other systems.
Section VIII summarizes proposed approaches that focus on the design of a communication system to facilitate spectrum sharing.
Section IX describes cooperative approaches to spectrum sharing including the very active research areas of radar-communication system joint design and co-design as well as passive radar.
Section X discusses standardization efforts related to radar-communications frequency sharing.
Section XI compares approaches for spectrum sharing between radar and communication systems surveyed in this paper.
Lastly, Section XII summarizes the key findings and insights drawing conclusions about the state of radar-communications spectrum sharing and its future prospects.

II. CURRENT AND POTENTIAL DYNAMICALLY SHARED FREQUENCY BANDS USED BY FEDERAL RADARS
The United States (US) federal government is promoting the sharing of frequencies used by federal systems, including radar systems, as a way to promote US economic growth and technological leadership [6], [7], [8], [9], [10], [11].Several bands have been identified as candidates for sharing.Examples of bands that are used by federal radars and are initial candidates for spectrum sharing are listed below [2], [3], [12]  Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.
• 3500-3650 MHz (3550-3650 MHz is already shared with the CBRS): DoD ''operates high-powered shipborne, airborne, and ground-based radar systems in this band.These radar systems are used in conjunction with weapons control systems for the detection and tracking of airborne targets.The DoD operates radar systems used for fleet air defense, missile and gunfire control, bomb scoring, battlefield weapon locations, air traffic control, and range safety.This band is critical to military radar operations supporting national defense'' [3].

III. SURVEY PAPERS AND TAXONOMY OF DSS APPROACHES
This section provides an overview of key survey papers in the field of DSS along with a taxonomy of spectrum-sharing approaches.The survey in Figure 2 provides a large-scale categorization and analysis of various spectrum-sharing techniques and policies with an emphasis on radar and other wireless communication methods within the US while including more current references.This survey includes recent and past references, however, the past references are relevant to this survey and will be briefly discussed.Each of the prior surveys has specific goals and is used as a preliminary step for understanding more current research.Previous texts and survey papers summarize research related to spectrum sharing between radar and communication systems.Topics covered include general approaches to dynamic spectrum sharing and cognitive radio (CR) communications [13], [14], [15], [16], [17], [18], [19].Topics also include spectrum engineering and management, including coexistence between radar and communication systems [20], [21], [22].
Regarding DSS generally, in 2005, Peha [15] addressed models for spectrum sharing and quality of service (QoS).Liang, et al. [16] present an overview of CR, a concept developed by Mitola and Maguire [23], including spectrum sensing techniques, cooperative spectrum sensing, cognitive spectrum access, MAC layer design impact of spectrum sensing, MAC protocols for CR, Cognitive Radio Networks (CRNs), routing and control of CR networks, emerging CR networks, economics of spectrum sharing, and challenges and open questions.Hayajneh and co-authors [17] focused on the coexistence of wireless body area networks (WBAN) with other wireless technologies.Bhattarai et al. [18] provides an overview of dynamic spectrum sharing and includes nearly 150 references.Spectrum access schemes and corresponding authorization regimes are identified, and examples in the US and elsewhere are described in [18].That paper also provides a timeline of spectrum initiatives.In addition, reference [18] overviews other topics including preventive and punitive approaches to reduce interference, coexistence of heterogeneous systems, privacy and security, spectrum management, and policy challenges.Doyle [13] describes Overlay and Underlay approaches to spectrum sharing, and Biglieri et al. [14] also describe the Interweaving approach.These and other approaches are described in Table 2. Traditional frequency reuse can be seen as a form of underlay in which frequencies are reused at a sufficient distance that aggregate interference received from the nearest co-channel transmitters does not constitute harmful interference.Similarly, transmitters in close proximity to one another are typically assigned different frequencies; out-ofband emissions of communications transmitters, although non-zero, are constrained by well-known methods such as using spectral masks to a level that does not cause harmful interference to nearby adjacent-channel systems.
More radar-centric aspects of DSS are considered in a survey by Griffiths and colleagues [20], which explains the problem of spectrum congestion from a radar perspective, and describes technical and regulatory approaches to solve the problem.Emerging approaches described include passive radar, waveform diversity, bioinspired design (e.g., based on the echolocation of bats and dolphins), and cognitive approaches.Labib and co-authors consider LTE-U (LTE for Unlicensed Spectrum), proposed to operate in the unlicensed 5GHz band occupied by radar and 802.11 technology (Wi-Fi).The Federal Communications Commission (FCC) has regulations on the sub-bands to enable fair spectrum sharing.
Figure 2 presents a comprehensive taxonomy encompassing various critical aspects of radar-communication spectrum sharing.This taxonomy categorizes key topics into four overarching domains: Spectrum Sharing Approaches, Sensing and Frequency Assignment Approaches, PU Operational Privacy, and System Design.Within these domains, several pertinent topics are explored, including Interwave, Overlay, Underlay, Frequency Hopping, Codewords, Co-design, Join, and Energy Consumption.This taxonomy serves as a foundational framework to guide discussions and research endeavors aimed at enhancing the coexistence of radar and communication systems within shared spectral environments.
Labib et al. surveyed RF regulations and radar systems in the sub-bands [21].Another survey by Labib and co-authors overviews radars in the 5 GHz and 3.5 GHz bands, spectrum awareness and dynamic spectrum access techniques, and spectrum sharing between communications and radar systems [22].Further, [22] describes cognitive radar as well as joint radar-communication system cognition, co-design, and cooperative operation as potentially useful technologies for sharing.Greco [19] presents a survey of cognitive radars, a technology envisioned by Haykin [24] that has the potential to enhance the coexistence between radar and communication systems and is discussed in more detail later in this document.
Based on the above resources and others discussed below, we present a taxonomy of DSS approaches in Table 2 and Table 3.

IV. DYNAMIC SPECTRUM SHARING (NOT SPECIFIC TO RADAR-COMMUNICATION SYSTEM SPECTRUM SHARING)
The effective utilization of channels in Radio Frequency (RF) is crucial to accommodate the growing number of wireless devices and alleviate congestion in crowded frequency ranges.Dynamic Spectrum Sharing (DSS) systems play a vital role in optimizing spectrum usage by sharing resources among various communication systems or users.This section explores key approaches in DSS that are not specific to radar-communication spectrum sharing.
The goal of spectrum sharing is to allow more wireless devices to operate over a larger variety of RF bands.Spectrum sharing can result in the general need to utilize underused frequencies and alleviate or free up overly congested bands.Spectrum sharing requires a series of protocols that allow wireless communication systems to operate using static band allocation.In a DSS system, resources are shared between different communication systems or users.These resources are dedicated to incumbent users but can be allocated to other non-incumbent users if they are not being used.The combination of various communication systems utilizing the same frequencies and prioritization of specific communication groups, including military and government radar as well as emergency and first responder organizations creates the need for Dynamic Spectrum Sharing (DSS).
Papers discussed in this section [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49] mainly focus on more detailed approaches rather than broad categories of methods for DSS, as described in Section III.These approaches are not presented as specific to DSS involving radars, although they may prove useful for radar-communications DSS.The tables presented in this section (Table 4 and Table 5) provide a comprehensive summary of various DSS approaches.Each row in the tables corresponds to a specific approach, with the corresponding survey paper for that topic noted in the first column.The remaining columns contain essential information about these approaches.The table includes references to the sources where these approaches are discussed, descriptions of the approaches themselves, their advantages, and known or potential disadvantages.This structured overview will aid readers in quickly grasping the key characteristics and considerations associated with different DSS techniques.
Table 6 provides a summary of various DSS approaches with their respective advantages and disadvantages.These approaches include dedicated spectrum control channels, listen-before-talk strategies, channel selection based on rewards, chance-constrained programming, frequency hopping, and power control.Each approach addresses the challenge of efficient spectrum utilization in different ways.For instance, dedicated control channels offer straightforward implementation but may become congested, while frequency hopping allows for handling more traffic but introduces overhead with frequent channel changes.These strategies showcase the diverse range of techniques employed to optimize spectrum sharing, each with its unique trade-offs, catering to different scenarios and requirements in the realm of wireless communication and radar systems.
A standard conceptual framework for DSS is illustrated in Figure 3.This architecture typically comprises components such as spectrum sensing, decision-making modules, communication devices, and, notably, a regulatory information database.Figure 3 additionally illustrates how these elements interact to facilitate real-time spectrum management and allocation, enabling the efficient sharing of available RF bands among diverse communication systems and users.Some DSS architectures are database-assisted, relying on aggregated information for various functions, including PU detection and Dynamic Spectrum Allocation (DSA).

V. SPECTRUM-SHARING POLICY
The United States FCC has created policies and procedures for spectrum sharing.The Wireless Innovation Forum (WIn-nForum) is an international non-profit organization that promotes the innovation of wireless communication including Software Defined Radio (SDR), CR, and DSA, in commercial and government sectors.Several executive discussions and articles published by the United States Department of Defense's Defense Advanced Research Projects Agency (DARPA) have described military applications for spectrum sharing and radar communications [50], [51].On an international level, spectrum sharing and wireless communications have an economic impact that will drive policy while trying to remain standardized with the WInnForum.This section provides an overview of key spectrum-sharing policies, the organizations driving these policies, and recent developments in this dynamic field.
Table 7 and Table 8 provide a chronological summary of key spectrum-sharing policies and documents in the United States.They include the publication year, central topic, responsible organization, and a brief summary of each policy's objectives or impact.These tables serve as a reference for understanding the evolution of spectrum management and the entities involved in shaping spectrum-sharing policies.
Recent years have witnessed significant developments in spectrum-sharing policy, as highlighted in Table 8.
Various academic, commercial, and military studies have considered spectrum occupancy, or the extent to which radio frequencies in various bands are occupied in time and space.
These studies have been summarized in Table 9.This includes studies by the National Telecommunications and Information Administration (NTIA), the US Department of Commerce agency that regulates and manages spectrum use by the federal government [62], [63], as well as studies by industry.Frequencies of interest include military radar, TV, and radio frequencies.The data reported may include the frequencies used, their power output, location, and time [62].

VI. DYNAMIC SPECTRUM ACCESS/SHARING APPROACHES
Within the broad categories of spectrum sharing identified earlier, several approaches have been proposed that are applicable, but not specific, to sharing between communications and radar systems.Some spectrum-sharing architectures focus on sharing frequencies that are heavily used by primary or incumbent users and therefore can only be accessed by licensed SUs on a case-by-case opportunistic basis.Spectrum access systems require the ability to sense the presence of a PU and often crowdsource data or create contracts with SUs to act as spectrum-sensing nodes.In military applications, e.g., involving a radar mounted on a Naval ship, other spectrum users should not be aware of the radar's precise location.In these instances, the granting protocols may have 138352 VOLUME 11, 2023 Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.obscuring features that prevent other users from discovering the presence and location of a PU.These approaches can be done by changing the approval process of spectrum grants to hide the PU information from non-priority users.

A. CENTRALIZED/DATABASE-ASSISTED
Centralized/database-assisted spectrum access is used for spectrum sharing in the Citizens Broadband Radio Service (CBRS) and the TV white space (TVWS) at 470-790 MHz, which shares the 3550-3650 MHz band with government radars.CBRS employs central coordination using a spectrum access system (SAS) to assign frequencies through spectrum grants in response to requests by CBSDs (CBRS devices) such as 5G base stations.The CBSDs, in turn, control frequency use by their client devices, e.g., user equipment (UE).The SAS architecture is database-assisted with a common core that aggregates all of the data.The SAS serves as a cloud-based computing device to make decisions on frequency assignment.The system collects all of the data points for spectrum sensing and grant requests and makes decisions regarding the approval of spectrum grants accordingly.The SAS performs a function similar to the TVWS database used to coordinate spectrum sharing in US television broadcast bands and the automated frequency coordination (AFC) planned for spectrum sharing in the US 6 GHz band [25].In contrast to TVWS spectrum sharing, which permits autonomous sensing as an alternative to the use of the database, CBSDs must be authorized by the SAS to transmit; CBSDs also report results of their sensing to the SAS [74], Part 96.
Table 11 summarizes key references related to centralized /database-assisted spectrum access and sharing methods.These references explore topics such as base station clustering, beamforming, database-aided sensing for radar bands, and the SAS architecture for CBRS.They offer insights into the design and evaluation of these approaches, enhancing understanding of centralized spectrum management techniques.

B. OPPORTUNISTIC SHARING
In both cooperative and non-cooperative scenarios, opportunistic sharing focuses on allocating spectrum resources in a situation where a PU frequently or occasionally transmits, leaving SUs with intermittent bursts of opportunity to use the designated frequencies.This approach was adopted for 5 GHz U-NII band dynamic frequency selection (DFS) and can be used in the TVWS band when required [25] and [74] (FCC Rules, 47 CFR, Part 15, Subpart H).
Table 12 summarizes relevant references focusing on opportunistic spectrum sharing.These studies explore various aspects of sharing spectrum with rotating radars and cellular networks.By examining these references, one can gain insights into the methods, results, and conclusions related to opportunistic spectrum sharing in different scenarios.

C. SENSING APPROACHES: COOPERATIVE SENSING, ETC
To accurately detect the presence of a PU, a SAS will use resources such as Environmental Sensing Capability (ESC), crowdsourced SU-sensed spectrum occupancy data, and potentially information supplied by PUs to sense the surrounding areas.Many approaches focus on crowdsourcing information and can be applied to centralized and decentralized systems.The following cooperative sensing techniques assume the data collected by the sensing nodes in the radio environment to be true.The ways in which they are used denote several different data collection types and their respective utilization methods.Figure 5 shows a system that is informed of a PU's transmission activity directly by PU.In contrast, Figure 4 shows a system that must use some sort of spectrum sensing to detect a PU's transmission activity.Table 13 describes various methods for spectrum sensing.While the approaches differ in their methods of identification, their primary objective is to accurately detect an actively transmitting PU.

D. PU PRIVACY PROTECTION
There are two main requirements for protecting PUs: protecting the ability of their RF systems to function by reducing interference and protecting their privacy.In the case of military radars and other military PUs, protecting PU identity and operational information such as location, travel path, and spectrum usage is highly desirable.Interference from SU transmissions must be mitigated to protect PUs; therefore SAS architectures provide for notification of SUs to stop transmitting on frequencies used by PUs if PUs become active.The privacy aspect is more complex; several aspects including the location and channel of PUs' operation must be kept private and this can only be achieved if the SAS actively obscures pertinent information and/or relies on PU frequency occupancy/frequency availability information that is pre-processed in a way that obfuscates PU operational characteristics such as precise location, travel path, and frequency use.
Table 14 outlines various strategies employed to safeguard the privacy of PU within dynamic spectrum-sharing environments.These privacy protection approaches are instrumental in preventing SUs or malicious entities from gaining detailed insights into PU activities, locations, or frequency allocations,   Table 14 offers a concise exploration of five key privacy protection techniques, along with their respective references, descriptions, advantages, and disadvantages.These strategies encompass methods such as frequency hopping, obscuring spectrum usage through randomization, codeword-based encoding, energy consumption, and output management, and the introduction of false PU information into databases.Understanding the nuances of these privacy-enhancing mechanisms is crucial in designing robust and secure dynamic spectrum-sharing systems.
Protecting the privacy of PUs in spectrum access systems has a strong impact on the quality of service to SUs.Creating false entries into a SAS or REM's database and assigning SUs frequencies in a way that avoids the actual frequencies occupied by the PUs, with just enough difference to obscure the information that could be maliciously collected and interpreted to infer PU operational details such as location, velocity, and frequency use shows to be a strong candidate for obscuring PUs in database-backed SASs.Randomization of assignments and false entries is a common solution proposed by researchers trying to obscure the characteristics of a PU.An issue with randomization of SU allocation and false PU usage is with randomness itself.If the PU's frequency usage results in a pattern in the apparently random information generated by the system, the SUs could find the pattern using several statistical analysis methods.The act of randomizing unfortunately leaves additional and unnecessary frequency ranges unused.If a system is going to falsely emulate a PU to obscure a true PU, the system must do what it would do for a true PU.Unless the scheduling algorithm synthetically tells SUs to vacate the channel in the middle of their grants, PU usage could be detected from any grant suspension.A solution would be to have both randomness and overlapping false schedules that try to limit the amount of spectrum falsely occupied to provide both PUs and SUs with the best quality of service.The more layers of obscurity, the less spectrum will be available to SUs.The more the false PU information differs from the true PU, the easier it is for malicious users to determine what is real and what is fake.False location data needs to be adjacent to the actual PU data.If a false PU is logged far away from the actual PU, any device that can receive transmit power can easily determine whether the data corresponding to the PU's actual operation is accurate.
Future work in PU privacy should focus on obscuring the PU mission and identity, the PU location, and the specifics of spectrum utilization.Obscuring PU location has the least impact on spectrum availability for SUs.Algorithms for PU privacy protection should focus on truly emulating PUs and creating entries similar to those for the actual PUs so false entries cannot be easily detected.Protecting the identity of the PU is still a current issue not directly addressed.Identity protection will have to be accomplished indirectly through encrypting transmitted data and obscuring spectrum use and location.

VII. RADAR-CENTRIC APPROACHES TO ENHANCE SPECTRUM SHARING: SIGNAL PROCESSING AND COGNITIVE RADAR
This section focuses on aspects of radar design that enable radars to operate in the presence of interference or avoid interfering with communication systems, without prior knowledge of the communication systems.Signal processing approaches including various MIMO and beamforming techniques have been proposed, as well as approaches that focus on radar waveform design and adaptation of the radar, including the concept of CR [24].

A. RADAR SIGNAL PROCESSING AND MULTI-INPUT MULTI-OUTPUT
Table 15 summarizes various signal processing approaches, with a focus on MIMO techniques.These methods harness advancements in machine learning, spectrum sensing, transmit notching, null-space projection, and other specialized techniques.Each approach is evaluated based on its advantages and disadvantages, providing an overview of radar signal processing strategies for enhanced spectrum sharing.
138358 VOLUME 11, 2023 Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.by Haykin [24], aims to manipulate radar operation to optimize target detection while minimizing perceived interference from and/or to local communication transmitters and receivers.Waveform design, detailed as radar waves, may have flexibility in certain characteristics (e.g.frequency, direction, power) and still deliver similar results.
Table 16 offers insights into different cognitive radar approaches, including null space radar projection, adaptive radar for maximum Signal to Noise Ratio (SNR), optimized target detection, calculated band selection, and static waveform design.Each approach's advantages and disadvantages are highlighted to provide a comprehensive understanding of their utility in spectrum-sharing scenarios.

VIII. COMMUNICATION SYSTEM-CENTRIC APPROACHES
Other research, e.g., [137], [138], [139], [140], and [141], has focused on interference to secondary-user communications systems and design of these systems to mitigate interference from PUs such as government radars.Representative publications are summarized in Table 17.Overall, systemcentric approaches could apply to providing spectrum occupancy data to PUs and spectrum operators.Systemcentric architectures may be slower or have more overhead processing than other architectures, but these models are not evaluated on speed.

IX. RADAR & COMMUNICATION SYSTEM CO-DESIGN/JOINT DESIGN AND PASSIVE RADAR
Radar and communication systems can be designed jointly to ensure coexistence during operation, or replaced by integrated, dual-or multi-function systems that perform both communications and radiolocation.Researchers have proposed several strategies for creating heterogeneous systems.Initially, Joint radar-communications systems designs 138360 VOLUME 11, 2023 Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.have been suggested with the intention of maximizing hardware utilization and minimizing costs [142], [143].When permitted by regulatory and operational constraints, joint design of radar and communication systems can be an efficient way to share the spectrum between these two types of applications.
Figure 7 illustrates a concept of co-design between radar and communication systems.The diagram features a shared infrastructure represented by a tower, where both radar and communication nodes coexist and collaborate seamlessly.Atop the tower, there is an aerial vehicle representing a radar node, while wireless devices are dispersed on the ground or closer to ground level.The wireless devices may be fixed in place or mobile devices.

X. REGULATIONS AND STANDARDIZATION
Regulations such as Part 15, Subpart H and Part 96 of the FCC rules [74] establish requirements for radio devices and Spectrum Systems (SASs) for use in a shared spectrum.Specifically, Part 96 establishes and regulates the CBRS within the 3.5 GHz band shared with federal radars.However, more detailed specifications are required to ensure consistent functionality and interoperability of spectrum-sharing systems.The Spectrum Sharing Committee (SSC) within the Wireless Innovation Forum, an industry organization whose members have interests related to dynamic spectrum sharing, cognitive radio, and softwaredefined radio, is developing and maintaining a set of standards for CBRS-SAS, CBSDs, SAS-CBSD, and SAS-SAS interfaces, and other protocols and features of DSA implementation [215].
Table 20 highlights three pivotal standardizations that govern spectrum sharing between radar communication systems.Although these standardizations may undergo future expansions or enhancements, they currently provide fundamental policies and descriptions that shape the relationships and interactions between radar and communication systems within shared frequency bands.

XI. COMPARISONS
The topics surveyed in this paper have been summarized in the previous sections.Our comparisons and characteristics based on our findings are summarized in this section.
Figure 8 indicates the surveyed effect or relationship each approach has on the other.''Not Compatible'' indicates that the two systems are not typically compatible with each other or function with opposing topics, e.g.Cooperative Systems and Non-Cooperative Systems would be designed using conflicting sources of information and would not make sense to be included in the same architecture.''Hinders/Non-Ideal'' refers to a non-ideal combination that in some instances could hinder the accuracy, compliance, or performance of the affected system.''Enhances/Benefits'' directly refers to an approach that is able to increase flexibility, accuracy, compliance, robustness, or maybe a critical component of the affected system.''No Effect'' in the figure is defined as not having a positive or negative effect and are not opposing approaches.''Includes/Included In'' is a two-way relationship where the approach is an approach containing another approach or vis versa.The ''Same Approach'' tag indicates that the two approaches are the same.
The surveyed approaches all contained six main overarching characteristics that were found to be important for proper compliance and functionality.The approaches include flexibility, simplicity, PU interference protection, robustness, PU privacy, and efficiency.
• Flexibility refers to the approach's need for the ability to handle atypical events and information or the ability to be implemented with or within various systems.
• Simplicity refers to the need for the approach solution to be easy to implement or relatively straightforward in its algorithms or implementation.Most approaches were not found to have a great need for simplicity.
• PU Interference Protection refers to the requirement of the approach to be non-interfering with incumbent users or military radar.Most values for this category were scored highly based on FCC requirements.
• Robustness is described as the approaches needing to be adverse to rigorous input or use.PU privacy relates to the required need to protect the identity of the incumbent communication system.This is necessary for most approaches, but in the case of Non-Cooperative Systems and Signal Processing, they are relatively independent of that trait.
• Efficiency refers to the ability of the system to be efficient and has minimal wasted computation or use.This characteristic is relatively arbitrary, but for some system structures, like centralized database-assisted and 138364 VOLUME 11, 2023 Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.cognitive radar, it is important to be efficient to ensure an expected level of functionality and compliance.
DSS does not need to be simplistic for proper function.DSS does require robustness in order to prevent PU Interference but does not need to necessarily protect PU privacy or be efficient.Radar-centric communication must be robust in order to prevent interference with PU activity.Radar and Communication-Centric and Radar/System-Centric Co-Design must have flexibility in order to handle multiple types of communication but do not necessarily need to be robust or protect a PU.Signal Processing must be robust to properly process transmissions, but functionality is not typically dependent on efficiency or simplicity.Signal processing has nothing to do with PU interference protection or PU privacy.Cognitive communication must prevent PU Interference Protection and must be robust to function.Noncooperative systems must be robust and have a primary goal of preventing PU interference.Centralized database-assisted designs must be robust to prevent interference and require some level of PU privacy protection and efficiency in order to perform their tasks.Communication-centric approaches must be efficient and protect PU privacy.Cooperative systems must be robust, prevent PU interference, and protect PU privacy to be permitted to function.Cognitive radar must also protect PU privacy and prevent interference while remaining robust enough to handle various architectures.
These analyses are based on the collected data in previous sections.While the results are subject to refinement as specific implementations are considered, they are used 138366 VOLUME 11, 2023 Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.to identify key features and needs between each of the approaches.

XII. CONCLUSION
Concluding Thoughts: As spectrum sharing between radar and communication systems continues to grow in ubiquity, we expect the demand for both informed and uninformed relationships to form between various communication and radar systems to grow.To date, radar-communications spectrum sharing in the United States has focused on the 3.5 GHz CBRS band.Because spectrum sharing regulations and technology as well as legacy systems are well established in the CBRS band, it is more likely that any substantial innovations will be implemented in bands that are yet to be shared and therefore open to novel approaches.Research relevant to radar-communications spectrum sharing has been underway for at least 20 years.Although major, disruptive innovations cannot be ruled out, we anticipate that future advancements in radar-communications spectrum sharing will embody variations, combinations, derivatives, or enhancements of the general methodologies identified in the preceding sections.Our thoughts on each major topic are summarized below.
Spectrum Policy and Occupancy: The spectrum policies and occupancy studies provide protocols, guidelines, and availability of the radar spectrum.In various studies of different bands, it is shown that there is low spectrum utilization.Most notably, the 3.5 GHz Maritime Radar Band was surveyed in 2014 and shows low high-strength radar usage.
Radar-communication and System Co-design: Radarcommunication system co-design and similar approaches may be useful for the coexistence of heterogeneous government systems such as government radars and government communication systems.However, this approach is unlikely to be viable for coexistence between federal radars and commercial communication systems unless (1) the flow of information is only from communication systems to the radars and (2) methods are devised to mitigate risks to radar privacy, e.g., through systematic manipulation of the communication systems that cause the radars to adapt in a way that reveals unintended information about the radars' operations.
Incumbent Informed DSS: Regarding a DSS system for use in the US, if cost and resources allow, an incumbentinformed solution would be ideal.If a DSS system operator can provide a strong link between DoD entities, commercial carriers, and possibly unlicensed users, it has the greatest chance of success.Approaches identified in the literature present a range of trade-offs between increased spectrumsharing efficiency and PU privacy.
Data Format: NATO currently uses XML for data formatting so it would be recommended to use NATO's XMLbased standard if interaction with NATO forces is anticipated [216].JavaScript Object Notation (JSON) is another option for data formatting.All data should be encrypted and

A. FUTURE WORK
Based on various papers surveyed, we believe that spectrum sharing between radar and communications will be uninformed and use sensing models for PU transmission detection.Due to the uncertainty of using crowdsourced spectrum sensing data, most spectrum sensing data will be collected from operator-owned nodes implemented for private networks or existing large-area networks like those implemented by cellular providers.
Advancements in PU obfuscation will most likely appear in US military or federal implementations of communication networks where PU interference protection and privacy are the main concerns.We do not foresee any form of spectrum reallocation which will cause more adaptations of spectrum sensing.
Open-source frameworks for spectrum sharing and CR have already begun to spark interest and development into more applications of PU detection.Based on more current research, we expect to see an increase in spectrum utilization across the US.These increases in the utilization of available spectrum will mainly be in the CBRS but will expand into other frequency ranges such as those for deep space observation and small private networks.
Future improvements in spectrum sharing between radar and communications will most likely lead to an increase in the number of smart cities where various types of spectrum sharing, PU detection, will be utilized by both public and private sectors [217].
Reconfigurable intelligent surfaces (RISs) are programmable surfaces that can control the reflection of electromagnetic waves [218], [219], [220].As research enhances RISs, we expect to see it become part of network installations and key components in DSS where detecting if a PU is actively transmitting is necessary to a system that is not directly informed by the PUs themselves.RISs will enhance wireless communication, and the application will begin to appear in areas of high wireless usage and environments that heavily obstruct wireless signals such as densely populated urban centers with towering buildings.
This survey primarily focuses on the policies and approaches of spectrum sharing between radar and communications within the US.While all policies considered are US-based, many of these techniques are applicable to other regions of the world.Regulations and permissions can vary based on the frequency bands, the types of equipment or systems involved, and the specific use cases.These techniques should only be implemented if they are permitted in the country or governing region of operation.We expect many of these approaches to be useful worldwide for spectrum sharing in both radar and communications, whether used together or separately.
In conclusion, the future of worldwide spectrum sharing between radar and communications is poised for nonincumbent-informed sharing based on sensing models, with data predominantly collected from operator-owned nodes in private and large-area networks.The advancement of PU obfuscation will predominantly occur in US military and federal communication networks, prioritizing PU privacy.Opensource frameworks for spectrum sharing and CR are fueling the development of new PU detection applications, leading to increased spectrum utilization across the US.Looking ahead, this trend is expected to result in more smart cities adopting spectrum sharing and PU detection, with a continued emphasis on improving cooperation and coexistence between radar and communications systems.As research progresses, the collaboration between these domains will likely drive further innovations in spectrum-sharing techniques and lead to more efficient utilization of available wireless resources.

FIGURE 1 .
FIGURE 1. Venn diagram of topics related to spectrum sharing between radar and wireless communication systems.

FIGURE 2 .
FIGURE 2. Taxonomy of topics included in this survey.

FIGURE 4 .
FIGURE 4. Cooperative sensing model uninformed by PU. (No line of communication between incumbent and fusion center.)

FIGURE 5 .
FIGURE 5. Incumbent informed DSS architecture.(Direct line of communication between incumbent and fusion center used to inform parties of PU activity.)

FIGURE 6 .
FIGURE 6. Multi-function system between different communication methods.

FIGURE 7 .
FIGURE 7. Radar and communication system co-design.

FIGURE 8 .
FIGURE 8. Compatibility and interaction matrix between various radar and communication approaches.The symbols indicate the effect of the approach from the corresponding row on the topic in the corresponding column.

FIGURE 9 .
FIGURE 9. Qualitative assessment of strengths of radar-communication system spectrum sharing approaches in terms of Efficiency, PU Privacy, PU Interference Protection, Flexibility, Simplicity, and Robustness.

TABLE 1 .
List of acronyms used in spectrum sharing, radar, and communications.
[12] perform missions critical to safe and reliable air traffic control (ATC) in the national airspace, border surveillance, early warning missile detection, and drug interdiction.These radar systems ensure the safe transportation of people and goods, encourage the flow of commerce, and provide for national defense.''[3].Col.Frederick Williams, USAF, acting director of the Office of Spectrum Policy and Programs, Office of the Secretary of Defense, quoted in[12], indicated that ''We're taking a look to determine if those federal agencies can develop a Spectrum Efficient National Surveillance Radar, while at the same time open that spectrum up for commercial broadband uses as well.''VOLUME11,2023138349

TABLE 2 .
Taxonomy of spectrum-sharing approaches.

TABLE 3 .
Taxonomy of spectrum-sharing approaches cont.

TABLE 4 .
Survey papers of dynamic spectrum sharing approaches: General.

TABLE 5 .
Survey papers of dynamic spectrum sharing approaches: Radar communications.

TABLE 6 .
Summary of dynamic spectrum sharing.
Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.

TABLE 9 .
Various spectrum occupancy studies.

TABLE 10 .
Dynamic spectrum access/sharing approaches.B. COGNITIVE RADAR/RADAR ADAPTATION/RADAR WAVEFORMS Advances in computing have fostered the development of autonomous systems.Adaptation loosely modeled on human cognition has proved feasible and reliable in robust software applications.Using cognition in radar has shown promise, and many researchers are exploring methods that best suit radar applications.Cognitive radar, a concept introduced

TABLE 14 .
Dynamic spectrum access/sharing approaches: PU privacy protection.

TABLE 15 .
Radar processing including MIMO and Beamforming.Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.

TABLE 17 .
Summary of communication system-centric approaches.Passive radars can be either bi-static or multi-static, use signals from existing communication systems to illuminate radar targets, and process the reflected signals to locate and those targets.

TABLE 19 .
Summary of passive radar studies.

TABLE 20 .
Rules and standards related to radar-communications spectrum sharing.