Communication, Sensing and Intelligent-System Applications

Introduction and Context

Nanoelectronics concerns the design and use of electronic devices whose critical dimensions, active materials, or functional layers are controlled at the nanometre scale. At this scale, quantum confinement, surface effects, tunnelling, ballistic or quasi-ballistic transport, and strong electrostatic coupling can substantially change device behaviour. These effects are not valuable merely because devices become smaller. Their importance lies in enabling new system capabilities: high-frequency operation, low-power sensing, mechanical flexibility, dense integration, and local intelligence.

This chapter examines how nanoelectronic devices and materials are used in communication, sensing, flexible electronics, the Internet of Things (IoT), edge artificial intelligence (AI), and energy-efficient intelligent hardware. The emphasis is on system-level design: selecting a device technology, matching it to an application, evaluating performance limits, and recognising reliability, security, environmental, and regulatory constraints.

A useful design principle is that no single nanoelectronic technology optimizes every requirement. A carbon-nanotube transistor may offer high transconductance and mechanical flexibility, while a silicon CMOS circuit may provide superior manufacturing maturity and integration. A two-dimensional material may provide excellent surface sensitivity, but its contact resistance and wafer-scale uniformity can limit practical deployment. Engineering judgement therefore requires a comparison of device physics, circuit architecture, packaging, algorithms, and application requirements.

Learning Objectives

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

1. Explain the role of nanoscale devices in RF, millimetre-wave, and terahertz communication systems.

2. Compare major nanoelectronic sensing mechanisms and evaluate sensor figures of merit.

3. Analyse flexible and wearable nanoelectronic systems, including their materials, interfaces, and reliability.

4. Design an IoT sensing node by considering energy, communication, computation, and security together.

5. Explain how edge AI is implemented using energy-efficient nanoelectronic hardware.

6. Evaluate a complete application using technical, economic, ethical, and regulatory criteria.

1. Nanoelectronic Platforms for Communication and Sensing

Important platforms include advanced silicon CMOS, silicon-germanium heterojunction bipolar technology, III–V compound semiconductors such as GaAs and InP, graphene, carbon nanotubes (CNTs), transition-metal dichalcogenides (TMDs), nanowires, ferroelectric materials, resistive-switching materials, and spintronic structures.

The appropriate platform depends on the system bottleneck. For a phased-array transceiver, gain, noise, linearity, phase accuracy, and thermal management are central. For a chemical sensor, surface functionalization, selectivity, drift, and calibration dominate. For a wearable node, mechanical strain tolerance, encapsulation, biocompatibility, and energy harvesting may be more important than maximum transistor speed.

At nanoscale dimensions, the surface-to-volume ratio increases markedly. This benefits sensors because adsorption of a small number of molecules can significantly change conductance, threshold voltage, capacitance, or contact resistance. The same feature can be a disadvantage in communication circuits because surface states, traps, and process variation can increase noise and instability.

2. Nanoelectronics for RF, Millimetre-Wave and Terahertz Systems

2.1 Frequency regimes and system requirements

RF systems commonly operate from hundreds of kilohertz to several gigahertz. Millimetre-wave systems generally occupy approximately 30–300 GHz, while terahertz systems are often described as operating from about 0.1 to 10 THz, although boundaries vary by application and community. As frequency increases, wavelength decreases according to

λ = c/f,

where λ is wavelength in metres, c is the speed of light in vacuum, approximately 3.00 × 10^8 m/s, and f is frequency in hertz. In a homogeneous dielectric, the phase velocity is lower than c and the effective wavelength is correspondingly shorter.

Short wavelengths permit compact antennas and large arrays in a small physical area. This makes beam steering and spatial multiplexing attractive at millimetre-wave frequencies. However, atmospheric absorption, blockage by objects or the human body, oscillator phase noise, packaging loss, and reduced transistor gain create substantial design challenges.

2.2 Device and circuit metrics

For a transistor used in a high-frequency amplifier, important metrics include cutoff frequency f_T, maximum oscillation frequency f_max, available gain, minimum noise figure, linearity, breakdown voltage, and power density. The cutoff frequency is the frequency at which the short-circuit current gain has magnitude unity. It is a small-signal metric and does not by itself guarantee useful power gain in a complete circuit.

For a field-effect transistor, a simplified long-channel estimate is

f_T ≈ g_m/(2πC_gg),

where g_m is transconductance in siemens and C_gg is the total gate capacitance in farads. The expression assumes a unilateral small-signal approximation and neglects parasitic resistances, delay, feedback capacitance, and distributed effects. In nanoscale and high-frequency devices, these neglected terms can be decisive; therefore measured or full equivalent-circuit values must be used for final design.

The noise factor is defined as

F = (SNR)_in/(SNR)_out,

where SNR denotes signal-to-noise ratio. Noise figure in decibels is NF = 10 log10(F). A lower noise figure is desirable. In a receiver, low-noise design requires careful control of gate resistance, channel noise, flicker noise, substrate coupling, lossy interconnects, and impedance matching.

For a thermally limited resistor, the voltage-noise power spectral density is

S_v(f) = 4kTR,

where k is Boltzmann’s constant, T is absolute temperature, and R is resistance. This relation describes white Johnson–Nyquist noise under the usual lumped, equilibrium assumptions. Nanoelectronic devices also exhibit shot noise, generation–recombination noise, random telegraph noise, and 1/f noise. At low frequencies, 1/f noise and individual charge traps can dominate over thermal noise.

2.3 Antennas, arrays and beamforming

A single antenna at millimetre-wave frequency may have limited coverage and can be blocked easily. Arrays overcome this limitation by combining signals from multiple antenna elements. In an ideal coherent array, the desired direction experiences constructive interference when element phases are aligned. Practical arrays suffer from mutual coupling, amplitude and phase errors, finite-resolution phase shifters, calibration drift, and power consumption in the many RF chains.

Hybrid beamforming divides the task between an analog network and a smaller number of digital baseband chains. This reduces converter and processor power, but it restricts the number of independent spatial streams and can complicate calibration. Nanoelectronic phase-change devices, memristors, tunable capacitors, and compact RF switches are being investigated for reconfigurable beamforming networks.

2.4 Terahertz opportunities and limitations

Terahertz communication offers potentially very large bandwidth for short-range links, chip-to-chip interconnects, imaging, spectroscopy, and high-resolution radar. Yet semiconductor gain decreases with frequency, atmospheric attenuation can be severe at molecular absorption lines, and efficient sources and detectors remain difficult to realize. Graphene and other two-dimensional materials are attractive because of high carrier mobility in suitable conditions, tunable conductivity, plasmonic confinement, and atomically thin form factors. Their practical performance depends strongly on substrate, defects, contacts, dielectric environment, and fabrication uniformity.

A sound terahertz design therefore begins with a link budget rather than with a material claim. The budget must include spreading loss, atmospheric absorption, antenna gain, mismatch, packaging, phase noise, receiver noise, and required modulation margin. Laboratory demonstrations at very short distances should not be interpreted as evidence of immediate large-scale deployment.

3. Nanoelectronic Sensors

3.1 Sensing chain

A complete sensor is more than a sensitive nanomaterial. It consists of a recognition mechanism, transducer, readout circuit, signal-conditioning stage, calibration model, communication interface, and decision algorithm. Errors may arise from the sensing element, temperature, humidity, mechanical stress, electronics, or data interpretation.

The sensitivity of a measurement can be expressed locally as

S = Δy/Δx,

where x is the measurand, y is the sensor output, and Δ denotes a small change over the operating region. The units depend on the output variable. Sensitivity alone is insufficient: a high sensitivity sensor may have poor selectivity, high noise, or severe drift.

The limit of detection is often estimated as

LOD ≈ kσ/S,

where σ is the standard deviation of the baseline signal, S is sensitivity, and k is a selected decision factor, commonly around 3 for an informal detection estimate or a larger value when statistical confidence and false-alarm requirements are specified. This equation assumes approximately stationary noise and a locally linear response. It should not be used without considering matrix effects, calibration uncertainty, and non-Gaussian noise.

3.2 Electrical and chemical nanosensors

Nanowire and CNT field-effect sensors detect changes in conductance or threshold voltage caused by surface charge, adsorption, or electrochemical reactions. Graphene and TMDs offer large exposed surfaces and can be functionalized with receptors such as antibodies, enzymes, aptamers, or molecularly imprinted polymers. Nanopore sensors infer molecular properties as charged molecules pass through a nanoscale pore and modulate ionic current.

Electrochemical sensors use potential, current, impedance, or charge as the measurement variable. In a simplified faradaic process, current is related to reaction rate and analyte concentration, but the exact response depends on electrode geometry, diffusion, kinetics, supporting electrolyte, and interfering species. Consequently, calibration in the intended sample matrix is essential.

Biosensors require special attention to biofouling, nonspecific binding, sterilization, storage, and biocompatibility. A sensor that performs well in buffered laboratory solutions may fail in blood, saliva, wastewater, or soil. Surface chemistry and packaging are often as important as the active nanomaterial.

3.3 Physical sensors

Nanomechanical resonators, graphene membranes, nanowire beams, and piezoresistive structures can measure mass, pressure, force, acceleration, or biochemical binding. A resonator’s frequency shift can result from added mass or changes in stiffness. In simplified form, the resonance frequency of a mass–spring system is

f_0 = (1/2π)√(k_m/m_eff),

where k_m is effective stiffness in newtons per metre and m_eff is effective mass in kilograms. The relation assumes a linear, lightly damped oscillator. Actual nanomechanical devices require correction for mode shape, tension, damping, and environmental loading.

4. Flexible, Wearable and Printed Nanoelectronics

Flexible systems use substrates such as polyimide, polyethylene terephthalate, elastomers, paper, textiles, or ultrathin silicon. Conductors may be formed from nanowires, CNT networks, graphene, liquid metals, or printed metal inks. Organic semiconductors and oxide semiconductors can provide large-area, low-temperature fabrication, while transferred crystalline nanomembranes can provide higher performance.

Mechanical design must distinguish bending, stretching, twisting, and repeated cyclic deformation. A brittle conductor may tolerate bending around a large radius but fail under tensile strain. Serpentine interconnects, kirigami structures, neutral-axis engineering, encapsulation, and crack-arrest features improve durability.

Wearable biosystems typically include a sensor interface, analog front end, low-power processor, wireless link, power source, and protective package. Skin contact introduces sweat, motion artefacts, thermal variation, and privacy concerns. Data quality depends on mechanical attachment and user behaviour as much as on sensor resolution.

Energy harvesting can use photovoltaic, thermoelectric, piezoelectric, triboelectric, or radio-frequency sources. Harvested energy is variable and usually must be combined with energy storage and duty cycling. A practical design begins with an energy budget that includes quiescent current, sensing time, processing, transmission, leakage, and conversion losses.

5. IoT Architectures and Nanoelectronic Design

An IoT node usually contains sensing, local processing, memory, connectivity, power management, and security functions. The design objective is not simply minimum transistor energy. It is minimum energy per useful decision while maintaining accuracy, latency, lifetime, and trustworthiness.

Communication often dominates energy consumption. Local feature extraction or event detection can reduce the number of transmitted bits. However, computation is not automatically cheaper than communication; the comparison depends on processor architecture, radio distance, protocol overhead, and workload. Adaptive duty cycling, wake-up receivers, compressed sensing, event-driven acquisition, and federated learning can reduce energy and protect raw data.

Nanoelectronic nonvolatile memories, including resistive RAM, phase-change memory, magnetic RAM, and ferroelectric memory, are relevant to IoT because they can retain information with low standby power. Important concerns include endurance, retention, variability, write energy, sneak paths, read disturbance, and device-to-device variation. Memory selection must consider the complete workload rather than a single headline density value.

Security must be designed into the node. Physical unclonable functions can exploit controlled manufacturing variation to generate device-specific responses. Hardware security modules, secure boot, authenticated updates, encryption, and key management are needed for connected deployments. A low-power device with weak update protection can become a large-scale security liability.

6. Edge AI and Intelligent Nanoelectronic Hardware

Edge AI places inference close to the sensor or data source. It reduces communication, latency, and exposure of raw data, but it is constrained by memory capacity, energy, thermal limits, and model accuracy. Common techniques include quantization, pruning, knowledge distillation, low-rank approximation, sparse computation, and approximate arithmetic.

A digital accelerator may use systolic arrays, single-instruction multiple-data units, or bit-serial arithmetic. Analog or mixed-signal in-memory computing performs part of a vector–matrix multiplication using conductances and currents. If conductance G_j represents a weight and voltage V_j represents an input, the ideal column current is

I = Σ_j G_jV_j.

This follows Ohm’s law and Kirchhoff’s current law for an ideal linear array. Real arrays introduce nonlinearity, parasitic resistance, limited conductance range, device noise, variation, sneak currents, analog-to-digital converter overhead, and programming error. Calibration and algorithm–hardware co-design are therefore essential.

Spiking neural hardware uses sparse events and stateful devices to emulate aspects of biological computation. Memristive, ferroelectric, phase-change, and spintronic devices may provide compact synaptic or neuronal functions. The research opportunity is substantial, but endurance, stochasticity, retention, learning stability, and manufacturing integration remain open problems.

Accuracy is only one evaluation metric. A useful edge-AI metric set includes energy per inference, latency, throughput, memory traffic, peak temperature, robustness to sensor drift, explainability, and performance under distribution shift. A model trained on laboratory data may fail when lighting, temperature, users, or biological conditions change.

7. Worked Example: Energy and Detection Limit of a Wearable IoT Node

Consider an illustrative wearable node that samples a chemical sensor once every 10 seconds. Each measurement requires 2 mJ for sensing and analog conversion, 0.5 mJ for local feature extraction, and 1.5 mJ for wireless transmission. Assume a 24-hour operating period and neglect sleep energy for a first estimate.

The number of measurements is

N = (24 × 60 × 60)/10 = 8640.

The energy per measurement is 2 + 0.5 + 1.5 = 4 mJ. Therefore,

E_day = 8640 × 4 mJ = 34,560 mJ = 34.56 J.

The average power is

P_avg = E_day/(24 × 3600 s) = 34.56/86,400 W = 0.0004 W = 0.4 mW.

This estimate excludes sleep current, regulator losses, retransmissions, display power, and battery self-discharge. If local processing can identify an event and reduce transmissions by 80%, the transmission contribution falls from 1.5 mJ to 0.3 mJ per sample, assuming the sensing and feature-extraction costs remain unchanged. The new daily energy is

8640 × (2 + 0.5 + 0.3) mJ = 24.192 J,

corresponding to an idealized average power of approximately 0.28 mW. The saving is significant, but it must be weighed against the energy and error cost of the local classifier.

Suppose a sensor has sensitivity S = 20 mV per micromolar and baseline standard deviation σ = 3 mV. Using the informal k = 3 estimate,

LOD ≈ 3(3 mV)/(20 mV/µM) = 0.45 µM.

This is a preliminary analytical estimate, not a validated clinical or environmental detection limit. Matrix interference, calibration bias, temperature dependence, and false-positive requirements must be tested experimentally.

8. Applied Case: Illustrative Smart Wound-Monitoring Patch

Background

A research and healthcare consortium proposes a flexible patch for monitoring wound temperature, pH, and selected biomarkers. The patch should alert a clinician to possible infection while minimizing patient discomfort and preserving privacy. This is an illustrative case rather than a report of a particular commercial product.

Problem

The system must operate on a curved, moving skin surface; tolerate sweat and mechanical deformation; distinguish clinically meaningful changes from motion and environmental effects; and operate for several days with limited battery capacity. Wireless transmission of continuous raw data is undesirable because of energy use and privacy risk.

Analysis

A nanostructured electrochemical interface offers high effective area and can be functionalized for selected biomarkers. A flexible oxide or organic transistor readout can be integrated on a polymer substrate, but drift and hysteresis require calibration. Temperature and pH compensation are necessary because electrochemical responses are often cross-sensitive. Mechanical strain near the wound can produce artefacts, so a strain reference element should be included.

The communication link should use encrypted, authenticated short-range transmission. Local processing can extract trends, rate of change, and confidence indicators. A clinician should not receive an unexplained binary alarm; the system should communicate the measured variables, quality flags, and decision rationale appropriate to the clinical workflow.

Approach and solution

The proposed architecture contains: (1) disposable sensing and fluid-interface layers; (2) a flexible analog front end with multiplexed electrochemical readout; (3) a low-power microcontroller with a compact anomaly-detection model; (4) nonvolatile memory for calibration coefficients; (5) intermittent wireless communication; and (6) a protected energy-storage element. The patch performs frequent local sampling but transmits summaries and alerts. A reference sensor and periodic calibration fluid or controlled baseline procedure are included where feasible.

Validation should proceed through benchtop testing, artificial wound-fluid testing, mechanical cycling, accelerated aging, small human-factors studies, and controlled clinical evaluation. Performance should be compared against a validated laboratory method. Model accuracy must be reported together with sensitivity, specificity, confidence intervals, missing-data handling, and subgroup performance.

Outcome criteria

Success is not defined by the smallest nanomaterial feature or the highest nominal sensitivity. It requires stable operation, clinically useful false-alarm rates, acceptable skin compatibility, secure data handling, reproducible manufacture, and an economically realistic disposable/reusable architecture.

Learning points

1. A sensor, circuit, algorithm, package, and clinical protocol form one system.

2. Local intelligence can reduce communication energy but introduces model-validation obligations.

3. Calibration and reference measurements are essential for nanosensors subject to drift.

4. Safety, privacy, and clinical usefulness are engineering requirements, not afterthoughts.

9. Challenges, Governance and Best Practices

Technical limitations

Nanoscale devices often show process variability, contact resistance, defect sensitivity, hysteresis, ageing, and environmental dependence. Laboratory performance may be measured on a small number of carefully selected devices and may not represent wafer-scale yield. Packaging can dominate the response of a sensor or the loss of an RF module. Thermal management becomes difficult when many devices are densely integrated.

Manufacturing and reliability

A credible technology assessment should report yield, repeatability, endurance, retention, temperature range, calibration burden, and failure modes. Accelerated testing must use physically meaningful stress conditions. Statistical process control and design-for-test are particularly important because nanoscale variability can affect both analogue and digital behaviour.

Ethical and regulatory issues

Wearable and biomedical systems collect sensitive information. Consent, data minimization, access control, secure storage, deletion policy, and secondary-use governance must be established. AI systems may produce unequal performance across demographic or physiological groups. Medical claims require appropriate regulatory pathways, traceability, risk management, and validation against accepted reference methods. Environmental assessment should include nanomaterial release, recycling, toxicity, and end-of-life disposal.

Best-practice design workflow

A robust workflow is to: define the application and failure consequences; specify measurable system requirements; select the sensing or communication principle; model energy, noise, thermal behaviour, and link performance; co-design device, circuit, package, and algorithm; prototype with reference instruments; characterize variation and ageing; test under realistic environments; secure the complete data path; and document uncertainty and limitations.

10. Contemporary Developments and Research Directions

Current research is moving toward heterogeneous integration, where mature CMOS is combined with two-dimensional materials, nanowires, photonic devices, ferroelectric layers, or neuromorphic elements. Chiplet and three-dimensional integration can shorten interconnects and improve system density, but thermal paths, test access, known-good-die assembly, and reliability become more complex.

Terahertz imaging and spectroscopy are progressing through improved sources, detectors, antennas, and computational reconstruction. Flexible and transient electronics are expanding into environmental monitoring and human–machine interfaces. Bioelectronic interfaces are becoming more multimodal, combining electrical, chemical, optical, and mechanical measurements.

In intelligent hardware, the major trend is co-design across device physics, memory, architecture, algorithms, and application requirements. The research challenge is not simply to demonstrate a novel switching device, but to establish reproducible system-level advantage under realistic workloads and manufacturing constraints.

Chapter Summary

Nanoelectronics enables communication and sensing systems through high-frequency devices, compact arrays, high-surface-area transducers, flexible materials, nonvolatile memories, and unconventional computing elements. RF and millimetre-wave systems benefit from compact arrays and heterogeneous semiconductor technologies, while terahertz systems offer bandwidth and sensing opportunities but face severe source, loss, and packaging limitations. Nanosensors can achieve high sensitivity, yet selectivity, drift, calibration, matrix effects, and reproducibility determine practical value. Flexible and wearable systems require coordinated mechanical, electrical, thermal, and biological design.

IoT nodes should be optimized for energy per useful decision rather than isolated device metrics. Edge AI reduces communication and latency but introduces memory, calibration, security, and model-validation requirements. In-memory and neuromorphic approaches are promising, but nonideal conductance, noise, variation, and endurance must be addressed. Successful applications integrate devices, circuits, packages, algorithms, communications, governance, and user needs. Professional evaluation therefore combines quantitative performance with reliability, privacy, safety, environmental responsibility, and manufacturability.

Key Terms

Nanoelectronics: Electronics using nanoscale structures, materials, or device dimensions.

Millimetre-wave: Electromagnetic frequencies approximately from 30 to 300 GHz.

Terahertz: A broad frequency region commonly spanning approximately 0.1 to 10 THz.

Cutoff frequency: Frequency at which a transistor’s short-circuit current gain reaches unity.

Noise figure: Degradation of signal-to-noise ratio expressed as a factor or in decibels.

Beamforming: Combining antenna signals with controlled amplitudes and phases to shape radiation.

Surface functionalization: Chemical modification of a surface to provide selective molecular interaction.

Sensitivity: Change in sensor output per unit change in the measurand.

Limit of detection: Smallest practically distinguishable analyte level under specified conditions.

Biofouling: Unwanted accumulation of biological material on a sensor or device surface.

Flexible electronics: Electronic systems designed to operate on bendable or deformable substrates.

IoT node: A connected device that senses, processes, communicates, and often actuates.

Duty cycling: Alternating active and low-power states to reduce average energy consumption.

Edge AI: Artificial-intelligence processing performed near the data source.

Quantization: Representing numerical values with reduced precision.

In-memory computing: Performing computation within or near the memory array.

Physical unclonable function: Hardware structure exploiting manufacturing variation for device identity or security.

Heterogeneous integration: Combining different materials, devices, or technologies within one system.

Distribution shift: Change between the data conditions used for training and those encountered in operation.

Design-for-test: Incorporating structures and methods that support manufacturing and field testing.

Review Questions

Short Answer Questions

1. Why does a high surface-to-volume ratio benefit nanoscale chemical sensors?

2. Define transistor cutoff frequency and explain why it is not a complete measure of RF performance.

3. What is the purpose of duty cycling in an IoT node?

4. State two practical limitations of analog in-memory computing.

5. Why are reference sensors useful in wearable biosystems?

Long Answer Questions

1. Explain the opportunities and limitations of nanoelectronics in millimetre-wave and terahertz communication systems.

2. Discuss the complete signal chain of a nanosensor, from molecular recognition to cloud or edge decision-making.

3. Compare the design requirements of a rigid RF transceiver and a flexible wearable sensor node.

4. Explain how edge AI can reduce IoT energy consumption and identify the associated risks.

5. Develop a validation strategy for a nanoelectronic biosensor intended for real-world deployment.

Analytical / Application Questions

1. A sensor has sensitivity 15 mV/µM and baseline standard deviation 2 mV. Estimate the informal detection limit using k = 3, and discuss why this value may not equal a validated field detection limit.

2. Design a high-level architecture for a battery-powered environmental IoT node that must operate for one year. Identify sensing, processing, communication, memory, energy, and security choices.

3. Evaluate whether a flexible nanomaterial sensor with excellent laboratory sensitivity should be commercialized. Construct a decision framework including selectivity, drift, yield, packaging, calibration, regulation, privacy, and environmental impact.

Multiple Choice Questions with Answers

Question 1

Which equation relates wavelength to frequency in free space?

A. λ = c/f

B. λ = cf

C. λ = f/c

D. λ = c + f

Correct Answer: A. λ = c/f

Question 2

What does transistor cutoff frequency f_T represent?

A. The frequency of maximum power dissipation

B. The frequency at which short-circuit current gain reaches unity

C. The frequency of zero gate capacitance

D. The frequency at which thermal noise vanishes

Correct Answer: B. The frequency at which short-circuit current gain reaches unity

Question 3

If a sensor sensitivity is 20 mV/µM and baseline standard deviation is 3 mV, the informal LOD for k = 3 is:

A. 0.15 µM

B. 0.30 µM

C. 0.45 µM

D. 1.80 µM

Correct Answer: C. 0.45 µM

Question 4

Which feature most directly reduces IoT communication energy by avoiding continuous raw-data transmission?

A. Local event detection

B. Higher antenna loss

C. Longer packet headers

D. Continuous maximum-rate sampling

Correct Answer: A. Local event detection

Question 5

In an ideal resistive in-memory array, the column current is represented by:

A. I = ΣG_j/V_j

B. I = ΣG_jV_j

C. I = ΣV_j/G_j

D. I = ΣG_j + V_j

Correct Answer: B. I = ΣG_jV_j

Question 6

Which factor is especially important when evaluating a biosensor in blood rather than a buffered laboratory solution?

A. Matrix interference

B. Vacuum wavelength only

C. Antenna beamwidth only

D. Transistor breakdown alone

Correct Answer: A. Matrix interference

Question 7

What is the principal purpose of beamforming in a millimetre-wave array?

A. To eliminate all atmospheric absorption

B. To control radiation direction through amplitude and phase combination

C. To make wavelength independent of frequency

D. To remove the need for calibration

Correct Answer: B. To control radiation direction through amplitude and phase combination

Question 8

Which statement about edge AI is most accurate?

A. It always consumes less energy than communication

B. It can reduce latency and transmitted data, but is constrained by memory and computation

C. It removes the need for model validation

D. It guarantees immunity to distribution shift

Correct Answer: B. It can reduce latency and transmitted data, but is constrained by memory and computation

Question 9

A wearable node uses 4 mJ per measurement and takes one measurement every 10 seconds. Ignoring sleep energy, its average power is approximately:

A. 0.04 mW

B. 0.4 mW

C. 4 mW

D. 40 mW

Correct Answer: B. 0.4 mW

Question 10

Which is a major governance concern for a smart biomedical patch?

A. Data privacy and clinical validation

B. Only antenna colour

C. Removal of all calibration procedures

D. Guaranteeing zero manufacturing variation

Correct Answer: A. Data privacy and clinical validation