Radar Systems Engineering, Governance and Future Development

Introduction and Context

Radar engineering at postgraduate level is not limited to waveform design, antenna selection, or detection theory. A deployable radar is a socio-technical system: it combines radio-frequency hardware, signal processing, software, communications, human operators, data infrastructure, regulation, safety controls, procurement arrangements, and long-term support. Its success is therefore determined not only by detection range or angular accuracy, but also by whether it satisfies an operational need, remains trustworthy under uncertainty, complies with law, can be maintained economically, and behaves safely in contested environments.

Systems engineering provides the disciplined process for converting mission needs into verifiable technical requirements and an operational capability. Governance provides the rules, authorities, accountability mechanisms, and ethical constraints within which that capability is acquired and used. Future radar development is increasingly shaped by digital arrays, software-defined architectures, distributed sensing, cognitive resource management, artificial intelligence, cloud-edge computing, and cyber-resilience.

This chapter presents an integrated framework for radar requirements definition, architecture selection, modelling, simulation, verification, validation, calibration, cybersecurity, safety, reliability, procurement, spectrum management, privacy, explainable artificial intelligence, and lifecycle development.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Translate operational needs into measurable radar system requirements.

2. Compare radar architectures using performance, risk, cost, regulatory, and lifecycle criteria.

3. Construct a model-based development and verification strategy.

4. Distinguish verification, validation, calibration, qualification, and acceptance testing.

5. Develop practical approaches to reliability, safety, cybersecurity, and configuration management.

6. Explain procurement, standards, spectrum, privacy, and ethical governance considerations.

7. Evaluate the role and limitations of artificial intelligence in radar systems.

8. Formulate a future-oriented digital radar roadmap.

1. From Mission Need to System Requirements

A radar requirement should describe an operationally meaningful capability under stated conditions. “High performance” is not a requirement; “detect a specified class of target with a stated probability at a stated range, in defined clutter and interference conditions” is closer to one.

A useful requirements hierarchy is:

• Mission requirements: the purpose of the system, such as airspace surveillance, maritime safety, weather observation, or automotive collision avoidance.
• Operational requirements: coverage volume, update rate, target classes, availability, deployment conditions, operator workload, and interoperability.
• System requirements: frequency band, transmitted power, antenna aperture, waveform agility, processing latency, detection performance, tracking accuracy, interfaces, and environmental limits.
• Subsystem requirements: receiver noise performance, exciter stability, array calibration accuracy, processor throughput, cooling capacity, and power quality.
• Verification requirements: the inspection, analysis, demonstration, or test by which compliance will be established.

Requirements must be complete, consistent, feasible, traceable, unambiguous, and testable. Each should identify a threshold value, a preferred objective where appropriate, operating conditions, and the verification method. A requirement that omits environmental assumptions can be misleading. For example, detection range depends on target radar cross section, propagation, clutter, interference, probability of detection, and false-alarm constraints.

A top-level operational concept describes who uses the radar, what information is produced, how quickly it is needed, what systems exchange data with it, and what happens when it fails. Stakeholder analysis is important because an air-traffic authority, military operator, spectrum regulator, maintenance organisation, privacy authority, and local community may have different legitimate concerns.

1.1 Requirement Decomposition and Traceability

A requirements traceability matrix links each stakeholder need to system requirements, design elements, test procedures, test results, risks, and accepted deviations. This matrix supports change control and prevents a common failure mode: a design feature being implemented without evidence that it satisfies an actual need, or a critical requirement having no objective verification method.

Requirements should also include quality attributes. Examples are availability, maintainability, cybersecurity, interoperability, usability, electromagnetic compatibility, energy consumption, and upgradeability. These attributes often interact. Increased waveform agility can improve electronic protection but increase software complexity and verification effort. Higher availability may require redundancy, which increases mass, cost, and maintenance burden.

2. Architecture Selection and Trade-Off Analysis

Radar architecture is the allocation of functions among antennas, transmitters, receivers, signal processors, data links, control systems, operators, and support infrastructure. Architecture selection should be based on a structured trade study rather than on a single performance metric.

Important alternatives include:

• mechanically scanned antennas versus electronically scanned arrays;
• passive electronically scanned arrays versus active electronically scanned arrays;
• monostatic versus bistatic or multistatic configurations;
• centralized versus distributed processing;
• dedicated hardware processing versus software-defined processing;
• single-sensor operation versus networked sensor fusion; and
• fixed installations versus mobile or spaceborne systems.

A decision matrix may score candidate architectures against weighted criteria such as coverage, detection performance, scalability, electromagnetic compatibility, cyber exposure, maintainability, cost, schedule, technological maturity, and regulatory feasibility. Scores should be accompanied by evidence and sensitivity analysis. If a small change in weights reverses the decision, the result is not robust and the stakeholders should discuss the underlying priorities.

2.1 System-Level Performance Reasoning

The radar range equation is a useful first-order model for a monostatic radar in free space:

\[

P_r=\frac{P_tG^2\lambda^2\sigma}{(4\pi)^3R^4L}

\]

where \(P_r\) is received power, \(P_t\) transmitted power, \(G\) antenna gain, \(\lambda\) wavelength, \(\sigma\) target radar cross section, \(R\) range, and \(L\) the product of losses represented as a dimensionless factor greater than or equal to one. The equation assumes far-field operation, a point target, matched polarisation, free-space spreading, and a simplified propagation model. It does not by itself determine detection probability.

Detection depends on signal-to-noise ratio, fluctuation statistics, the detector, integration strategy, clutter, interference, and the selected false-alarm probability. For thermal noise, a simplified receiver noise model is

\[

P_n=kT_0BF,

\]

where \(k\) is Boltzmann’s constant, \(T_0\) is reference temperature, \(B\) is receiver noise bandwidth, and \(F\) is the noise factor in linear units. This relation assumes a specified equivalent noise bandwidth and a receiver represented by a noise factor. It does not include clutter or external interference.

The \(R^{-4}\) dependence in the monostatic range equation is strategically important: doubling range requires a substantial increase in link-budget capability or a change in architecture, processing, deployment geometry, or target observability. A distributed or multistatic system may alter the geometric dependence, but it introduces synchronisation, association, communications, and calibration challenges.

2.2 Modelling and Simulation

Models should be used for specific decisions, not merely to generate attractive plots. A radar model may contain:

1. an environment model for propagation, terrain, weather, clutter, and interference;

2. a target and scattering model;

3. an RF and antenna model;

4. a signal-processing model;

5. a tracker and resource-management model;

6. a human or operational model; and

7. a cost, reliability, or support model.

Model fidelity should match the question. A low-order link budget is suitable for early feasibility analysis; it is not sufficient for validating detection performance in non-Gaussian sea clutter. Simulation data must be distinguished from measured data, and assumptions, random seeds, parameter ranges, and model versions should be recorded.

Digital engineering uses a shared system model containing requirements, interfaces, behaviours, data, risks, and verification evidence. Model-based systems engineering can expose interface conflicts early, but it does not eliminate the need for physical testing. A mathematically consistent model may still be invalid because its environmental assumptions are wrong.

3. Verification, Validation, Calibration, and Acceptance

These activities answer different questions:

• Verification: Was the system built according to its specified requirements?
• Validation: Does the system solve the intended operational problem in its real or representative context?
• Calibration: Are measurement channels and processing outputs related accurately to known references?
• Qualification: Can the design withstand specified environmental and stress conditions?
• Acceptance: Has the customer formally agreed that contractual requirements have been met?

Verification methods include inspection, analysis, demonstration, and test. A requirement such as equipment mass may be verified by inspection or measurement. A tracking performance requirement normally requires controlled testing with truth data. Validation may involve operational trials with representative users, realistic traffic, weather, clutter, and communications constraints.

A test plan should define configuration, instrumentation, reference standards, scenarios, sample size, environmental conditions, uncertainty, pass/fail criteria, and treatment of missing or anomalous data. Test results should preserve raw data and processing scripts. Statistical claims require attention to confidence intervals and independence assumptions. For example, many correlated measurements from one target track should not automatically be treated as many independent performance trials.

3.1 Calibration and Metrology

Calibration establishes the relationship between an instrument output and a reference value, together with measurement uncertainty. Radar calibration may address timing, frequency, amplitude, phase, antenna patterns, boresight, range, Doppler, polarisation, and channel-to-channel mismatch. In a digital array, calibration must compensate for temperature, ageing, mutual coupling, phase-shifter errors, receiver drift, and element failures.

Over-the-air calibration is valuable because it includes the complete RF path and antenna, but it is affected by propagation and reference-target uncertainty. Internal calibration improves repeatability and can run frequently, but it may not reveal antenna or radome effects. A sound programme combines factory calibration, installation calibration, operational monitoring, and recalibration after maintenance or major environmental change.

4. Reliability, Safety, and Lifecycle Management

Reliability is the probability that a system performs its required function for a stated time under stated conditions. Availability also includes repair and logistics. A system can have reliable components yet poor operational availability if repairs are slow or spares are unavailable.

For a simple series system, in which failure of any component causes system failure, and independent constant failure rates are assumed, the reliability is:

\[

R_s(t)=e^{-\left(\sum_i\lambda_i\right)t},

\]

where \(\lambda_i\) is the constant failure rate of component \(i\). This is an idealised model. Real radar systems may have repair, redundancy, common-cause failures, environmental dependencies, and non-exponential failure behaviour.

Reliability engineering uses failure-mode, effects, and criticality analysis; fault-tree analysis; accelerated environmental testing; preventive maintenance; built-in test; and life-cycle data. Redundancy must be analysed carefully: two identical units may share a power supply, clock, cooling system, software defect, or cyber vulnerability, so they are not fully independent.

Safety engineering identifies hazards, causes, consequences, controls, and residual risk. Radar hazards can include high-power RF exposure, rotating machinery, high-voltage supplies, lightning, electromagnetic interference with safety-critical equipment, incorrect target reports, uncontrolled antenna motion, and unsafe maintenance access. Safety cases should provide a structured argument supported by evidence. Safety is not equivalent to cybersecurity: a cyberattack may create a safety hazard, but safety analysis also covers non-malicious failures.

Lifecycle management includes concept, development, production, deployment, operation, maintenance, upgrades, retirement, and disposal. Configuration management must control hardware versions, firmware, software libraries, training data, models, test procedures, calibration files, and security credentials. Obsolescence planning is especially important for RF semiconductors, processors, operating systems, storage devices, and specialist test equipment.

5. Cybersecurity and Resilience

Modern radar attack surfaces include maintenance ports, network interfaces, navigation inputs, timing sources, software-update mechanisms, cloud services, supply chains, and AI data pipelines. Threats include unauthorised access, malware, denial of service, data manipulation, deceptive signals, spoofed time, malicious firmware, and extraction of sensitive operational information.

A risk-based cybersecurity programme should include asset identification, threat modelling, vulnerability assessment, security architecture, secure development, access control, cryptographic protection, logging, monitoring, incident response, recovery, and lessons learned. Security controls should support operational continuity. Excessive isolation can prevent necessary maintenance; excessive connectivity can expose critical functions.

Important design principles are least privilege, defence in depth, secure boot, signed updates, strong identity management, network segmentation, independent safety functions, fail-secure and fail-safe behaviour where appropriate, and verified recovery procedures. Supply-chain assurance should address provenance of components, software bills of materials, counterfeit detection, vulnerability notification, and vendor access.

Cyber resilience is broader than prevention. The system should detect abnormal behaviour, maintain essential functions under attack, degrade gracefully, restore trusted operation, and preserve forensic evidence. Radar-specific resilience also requires validation against spoofing, jamming, manipulated navigation data, corrupted tracks, and adversarial machine-learning inputs.

6. Procurement, Standards, and Spectrum Governance

Procurement should be capability-based and evidence-oriented. A statement of work should distinguish mandatory requirements, scored desirable features, government-furnished information, data rights, acceptance tests, training, spares, support, cybersecurity obligations, and upgrade responsibilities. Evaluation should consider total cost of ownership rather than acquisition price alone.

Open and well-defined interfaces reduce vendor lock-in and support incremental upgrades. However, “open architecture” is meaningful only when interface control documents, data rights, conformance tests, and governance responsibilities are clear. Intellectual-property arrangements should permit safety fixes, cybersecurity remediation, calibration support, and future interoperability.

Standards support repeatability and interoperability, but standards do not automatically guarantee performance or safety. Relevant categories include systems-engineering processes, quality management, environmental qualification, electromagnetic compatibility, information security, software assurance, safety management, and radar data exchange. The applicable standards depend on jurisdiction, mission, and sector; their current editions and contractual status must be checked rather than assumed.

Spectrum is a shared and regulated resource. Radar planning must consider national allocation tables, licensing, emission masks, out-of-band emissions, coordination zones, interference studies, coexistence, and international obligations. Spectrum access is not simply a technical choice of frequency. Adaptive or cognitive radar functions must remain bounded by authorisation and must not treat spectrum agility as permission to transmit arbitrarily.

7. Privacy, Ethics, and Responsible Surveillance

Radar data can reveal movement patterns, vehicle identities when combined with other data, household activity, or sensitive locations. Governance should therefore apply purpose limitation, data minimisation, retention limits, access control, auditability, lawful authority, and transparent accountability. Privacy risk depends not only on the sensor but also on fusion with cameras, communications records, databases, and location services.

Ethical surveillance requires necessity, proportionality, discrimination avoidance, human oversight, and mechanisms for challenge or redress where applicable. Automated classification should not be treated as ground truth. Errors may be systematic across environments, populations, target types, or weather conditions. Procurement contracts should require performance reporting by relevant operating condition, not merely one aggregate accuracy value.

Explainable AI is especially important when radar outputs influence safety, law enforcement, or defence decisions. Explanations may include salient waveform or range-Doppler features, confidence and uncertainty, comparable examples, model version, operating regime, and reasons for abstention. A visually plausible explanation is not necessarily a faithful explanation. Explainability should therefore be evaluated by whether it helps qualified users detect errors and make better decisions.

Human oversight must be substantive. An operator who can only accept an algorithmic recommendation without time, information, or authority to intervene does not provide meaningful oversight. AI systems should include confidence thresholds, out-of-distribution detection, abstention, logging, independent validation, and a defined response to model degradation.

8. Future Digital Radar Roadmaps

Future radar development is likely to be evolutionary rather than a single technological revolution. A practical roadmap can be organised into capability increments:

Increment 1—Digitised foundation: stable interfaces, accurate timing, data recording, calibration automation, modular software, and secure configuration management.

Increment 2—Networked sensing: shared tracks, distributed apertures, cooperative scheduling, common data models, and robust communications.

Increment 3—Adaptive operation: waveform and resource management that respond to mission priorities, interference, target uncertainty, and energy constraints within regulatory limits.

Increment 4—Assisted intelligence: AI for detection, classification, anomaly detection, maintenance prediction, clutter modelling, and operator decision support, supported by uncertainty estimates and human authority.

Increment 5—Resilient cognitive ecosystems: multi-sensor systems that learn from validated data, detect adversarial manipulation, preserve safety boundaries, and support rapid but controlled technology insertion.

Digital arrays can provide simultaneous beams, agile null steering, graceful degradation, and software-based mission changes. Their challenges include calibration scale, thermal management, data movement, power consumption, electromagnetic compatibility, software assurance, and complex failure modes. Edge processing reduces latency and communications burden, whereas central processing can simplify model management and fusion. Hybrid architectures are often preferable.

AI should be introduced through bounded use cases with measurable benefits. Predictive maintenance may be lower risk than autonomous target classification because maintenance recommendations can be reviewed before action. Nevertheless, maintenance AI can still create safety or availability consequences if its training data are biased or sensor health is misrepresented.

9. Illustrative Case: Coastal Integrated Surveillance Radar

Background

A fictional national maritime authority operates a coastal radar network for vessel safety, search-and-rescue support, and detection of unauthorised activity. The existing mechanically scanned radars have ageing transmitters, inconsistent calibration, and limited data exchange.

Problem

The authority proposes replacing them with high-performance digital radars. Initial procurement specifications focus on maximum range and antenna technology but omit availability, privacy, spectrum coordination, cyber resilience, and validation in heavy rain and sea clutter.

Analysis

A systems-engineering review identifies several risks. Maximum range is not operationally meaningful without target classes, radar cross sections, false-alarm limits, clutter conditions, and update requirements. A single high-power replacement may create a coverage gap during maintenance. Network connectivity increases the attack surface. Fusion with vessel-identification data introduces privacy and retention obligations. Finally, automated classification may produce unsafe confidence in small boats or unusual vessels.

Approach and Solution

The authority develops an operational concept and decomposes requirements into coverage, update rate, track continuity, availability, maintainability, cybersecurity, spectrum emissions, environmental tolerance, and operator workload. It compares three architectures: upgraded mechanical radar, active electronically scanned radar, and a hybrid network of moderate-performance digital radars with overlapping coverage.

The hybrid option is selected after lifecycle analysis. It provides graceful degradation, easier maintenance, and better coverage continuity, despite higher integration complexity. A digital engineering model links requirements to simulations, laboratory tests, sea trials, and acceptance evidence. Calibration uses internal references supplemented by periodic over-the-air measurements. Cybersecurity requirements include segmented networks, authenticated updates, monitored maintenance access, and recovery from trusted images.

AI is initially limited to operator decision support: clutter suppression recommendations, track-quality alerts, and maintenance anomaly detection. The system displays confidence, sensor provenance, and reasons for alerts. Automatic enforcement action is excluded from the initial release. Data retention and access policies are approved before operational deployment.

Outcome

The programme accepts a lower theoretical peak range than the original proposal but achieves higher network availability, better maintainability, clearer evidence of performance, and reduced single-point failure risk. Sea trials identify a classification weakness in unusual weather, leading to a restricted operating envelope and additional training data rather than an unsupported performance claim.

Learning Points

The case demonstrates that architecture quality is multidimensional. A radar should be judged by operational effect, evidence, resilience, compliance, and lifecycle value—not by a headline specification. Early governance decisions can prevent expensive redesign and reduce harmful overconfidence in automation.

10. Challenges and Best Practices

Major challenges include uncertain operational scenarios, incomplete environmental data, unrealistic laboratory testing, supplier dependence, software complexity, electromagnetic congestion, cyber threats, AI brittleness, and pressure to accept immature technology. Performance claims can also be distorted by selective test conditions or metrics that ignore false alarms, latency, track continuity, and operator workload.

Best practices include:

• begin with an operational concept and measurable mission outcomes;
• maintain bidirectional requirements traceability;
• use model fidelity appropriate to the decision;
• combine analysis, simulation, laboratory testing, and operational trials;
• quantify uncertainty and document assumptions;
• treat calibration and configuration as continuous lifecycle activities;
• analyse common-cause failures in redundant systems;
• integrate safety and cybersecurity from the architecture stage;
• include spectrum, privacy, and data governance in early design reviews;
• procure interfaces, data rights, training, and sustainment explicitly;
• introduce AI incrementally with independent validation and human authority; and
• plan retirement, disposal, and secure data destruction from the beginning.

Chapter Summary

Radar systems engineering converts mission needs into traceable, testable, and lifecycle-aware requirements. Architecture selection must balance performance with cost, risk, maintainability, interoperability, regulation, cybersecurity, and operational resilience. The radar range equation and receiver noise model support early reasoning but require clearly stated assumptions and cannot replace realistic detection and clutter analysis.

Verification establishes conformance, validation establishes operational suitability, calibration establishes measurement accuracy, qualification demonstrates environmental robustness, and acceptance records customer agreement. Reliability, safety, cybersecurity, configuration management, and obsolescence planning must be integrated rather than treated as late-stage activities.

Procurement should address total ownership, interfaces, evidence, data rights, and support. Spectrum governance, privacy, ethical surveillance, and explainable AI are central to responsible deployment. Future digital radars will increasingly use networked apertures, adaptive resources, edge computing, and AI, but these capabilities require bounded autonomy, uncertainty representation, secure updates, human oversight, and rigorous operational validation.

Key Terms

Requirements traceability: Linkage among needs, requirements, design elements, risks, and verification evidence.

Operational concept: Description of how users, systems, and organisations employ a capability.

Architecture: Allocation of functions and interfaces among system elements.

Radar cross section: Effective measure of a target’s ability to scatter radar energy toward a receiver.

Radar range equation: First-order relationship among transmitted power, antenna gain, wavelength, target cross section, range, and losses.

Verification: Demonstration that specified requirements have been met.

Validation: Demonstration that the system fulfils its intended operational purpose.

Calibration: Establishment of measurement relationships to known references with uncertainty.

Qualification: Evidence that a design withstands specified environmental or stress conditions.

Availability: Probability that a system is operational when required, including effects of repair and logistics.

Common-cause failure: Failure of multiple redundant elements due to a shared cause.

Safety case: Structured argument and evidence that a system is acceptably safe.

Cyber resilience: Ability to withstand, detect, respond to, recover from, and learn from cyber events.

Configuration management: Controlled identification and change management of system items and versions.

Spectrum governance: Regulation and coordination of radio-frequency use.

Data minimisation: Limiting collection and retention to what is necessary for an authorised purpose.

Explainable AI: AI whose outputs and limitations can be meaningfully interpreted by appropriate users.

Digital engineering: Use of connected models, data, and evidence across the system lifecycle.

Graceful degradation: Continued provision of reduced capability after failures or attacks.

Review Questions

Short Answer Questions

1. What characteristics make a radar requirement testable?

2. Distinguish verification from validation.

3. Why does the monostatic radar range equation contain an \(R^{-4}\) dependence?

4. What is the purpose of a requirements traceability matrix?

5. Name two privacy controls applicable to radar data.

Long Answer Questions

1. Explain how operational needs are decomposed into system and subsystem requirements, including the role of verification methods.

2. Compare mechanically scanned, electronically scanned, and distributed radar architectures using technical and lifecycle criteria.

3. Discuss the relationship among reliability, availability, redundancy, maintenance, and logistics.

4. Develop a governance framework for introducing AI-assisted radar classification into a safety-sensitive operation.

5. Explain why spectrum regulation, cybersecurity, privacy, and procurement should be considered during architecture definition rather than after design completion.

Analytical / Application Questions

1. A proposed radar claims twice the detection range of an existing system using the same target and propagation assumptions. Use the range equation to discuss the approximate required change in the combined link-budget terms, and identify why real performance may differ.

2. Design a verification and validation plan for a coastal radar network operating in rain, sea clutter, and heavy electromagnetic interference. Specify evidence, instrumentation, scenarios, and acceptance considerations.

3. Select between a single high-performance radar and several overlapping moderate-performance radars for a critical surveillance mission. Construct a decision framework that includes performance, availability, cybersecurity, spectrum, privacy, cost, and lifecycle risk.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes validation from verification?

A. Validation checks whether the system solves the intended operational problem; verification checks compliance with specified requirements.

B. Validation measures component failure rates; verification establishes spectrum licences.

C. Validation is performed only in software; verification is performed only in hardware.

D. Validation is customer acceptance; verification is always calibration.

Correct Answer: A. Validation checks whether the system solves the intended operational problem; verification checks compliance with specified requirements.

Question 2

In the simplified monostatic radar range equation, which variable has an inverse fourth-power effect on received power?

A. Wavelength

B. Target radar cross section

C. Range

D. Antenna gain

Correct Answer: C. Range

Question 3

What is the primary purpose of calibration?

A. To increase transmitted power beyond regulatory limits

B. To relate measurements to known references and quantify uncertainty

C. To replace operational validation

D. To eliminate all environmental effects

Correct Answer: B. To relate measurements to known references and quantify uncertainty

Question 4

Which condition is assumed by the simple series-system reliability expression using summed constant failure rates?

A. Perfectly dependent component failures

B. Independent components with constant failure rates

C. No component can fail during operation

D. Repair is instantaneous and unlimited

Correct Answer: B. Independent components with constant failure rates

Question 5

Which is an example of a common-cause failure?

A. One replaceable receiver module fails independently

B. Two redundant processors fail because they share a defective power supply

C. An operator performs preventive maintenance

D. A radar changes waveform according to a schedule

Correct Answer: B. Two redundant processors fail because they share a defective power supply

Question 6

Which procurement provision most directly reduces future vendor lock-in?

A. A single proprietary interface with no documentation

B. Open, controlled interfaces with conformance tests and appropriate data rights

C. A requirement for maximum peak power only

D. Elimination of acceptance testing

Correct Answer: B. Open, controlled interfaces with conformance tests and appropriate data rights

Question 7

Which principle limits radar data collection and retention to what is necessary for an authorised purpose?

A. Data minimisation

B. Graceful degradation

C. Beam steering

D. Fault-tree analysis

Correct Answer: A. Data minimisation

Question 8

Why is human oversight of AI not meaningful if operators cannot intervene or assess the recommendation?

A. Because radar signals cannot be digitised

B. Because nominal human presence does not provide substantive control or accountability

C. Because all AI outputs are necessarily incorrect

D. Because validation is unnecessary for automated systems

Correct Answer: B. Because nominal human presence does not provide substantive control or accountability

Question 9

Which activity is most directly associated with proving that a radar can withstand specified temperature and vibration conditions?

A. Qualification testing

B. Data minimisation

C. Spectrum allocation

D. Track fusion

Correct Answer: A. Qualification testing

Question 10

What is a major advantage of overlapping moderate-performance radars compared with one single radar?

A. They eliminate all calibration requirements

B. They can provide graceful degradation and reduce a single-point failure risk

C. They require no communications or synchronisation

D. They are automatically immune to cyberattack

Correct Answer: B. They can provide graceful degradation and reduce a single-point failure risk