Table 5 Existing privacy and security solutions in ILS (Part 1).

Federated learning (FL) for privacy and security^{57,59,79,87,98,99,112,113}

Federated learning for decentralized model training with differential privacy or transfer learning.

Protects privacy by not sharing sensitive data, improves scalability, and reduces data-sharing risks.

Challenges with large dataset scalability, high communication overhead, and limited labeled data for training.

FL enhances privacy-preserving localization while ensuring data accuracy and robustness across multiple devices. FL frameworks like FedLoc3D and FedPos improve accuracy and reduce communication overhead.

Vulnerable to model poisoning and data poisoning attacks where adversaries can inject false data to corrupt the model.

Crowdsourced localization, smart cities, healthcare, multi-building indoor navigation systems, location-based services.

Differential privacy (DP) for privacy preservation^{79,97,100,116,116,118}

Uses differential privacy to add noise to data and ensure individual privacy during localization.

Strong privacy protection, maintains system utility with noise addition, widely applicable in decentralized systems.

Balancing privacy and accuracy, especially when dealing with high levels of noise. Computational cost for large-scale systems.

DP ensures privacy in localization systems by using noise addition (e.g., Gaussian noise, local DP) to mask user data. It allows geo-indistinguishability for location privacy without significant degradation in query precision.

Attacks targeting the noise mechanism, such as reconstructing individual data from aggregate outputs, can reduce privacy.

Indoor location-based services, mobile crowdsensing, geospatial data privacy, privacy-preserving query systems, and healthcare.

Cryptographic techniques for secure localization^{60,64,114,115,121,122}

Cryptographic techniques (e.g., Paillier cryptosystem, homomorphic encryption) for securing location data during transmission and processing.

High level of confidentiality and security protects against unauthorized access or manipulation of data.

High computational cost and communication overhead, especially for large-scale systems. May not be scalable for real-time applications.

Secure cryptographic methods like homomorphic encryption and Paillier ensure confidentiality and prevent unauthorized access to sensitive location data. They can protect both user and service provider privacy.

Vulnerable to side-channel attacks and cryptanalysis, where attackers can exploit computational or transmission weaknesses.

Secure wireless positioning, IoT-based localization, secure mobile networks, cryptographically protected location-based services.

Blockchain for trust and security^89,123

Blockchain (e.g., Hyperledger Fabric) for providing immutable ledgers to authenticate and verify location data transactions.

Immutable ledger, increased trust and accountability, and provides transparency in location data transactions.

Scalability issues in large-scale environments, high energy consumption in consensus mechanisms, and integration complexity with existing systems.

Blockchain solutions ensure trust and security in localization systems by providing decentralized verification of location data. The use of permissioned blockchain (e.g., Hyperledger Fabric) addresses privacy concerns.

Susceptible to 51% attacks, where adversaries control the majority of the network and can manipulate the blockchain.

Secure navigation, supply chain tracking, transparent location-based data transactions, IoT, and data integrity in mobile and indoor localization systems.

Adversarial training and robustness^{82,83,85,90,120}

Adversarial training to improve system robustness by defending against attacks that manipulate sensor data or mislead models.

Improves model robustness, enhances resilience to adversarial attacks, and improves data integrity.

High computational cost, potential overfitting on adversarial examples, and scalability in real-time systems

Adversarial training techniques like label smoothing and feature squeezing improve the model’s resistance to adversarial inputs, even under low signal-to-noise ratio conditions.

Adversarial risks include adversarial examples designed to evade detection and fool the model, potentially causing mislocalization.

Robust indoor and outdoor localization, autonomous vehicles, security in AI-driven navigation, and defense against data manipulation attacks in wireless networks.

Privacy-preserving frameworks^5,81,116

Frameworks combining cryptography, anonymization (e.g., k-anonymity, l-diversity), and federated learning for privacy protection.

Comprehensive protection against unauthorized access, combines multiple privacy-preserving techniques.

Trade-off between privacy, accuracy, and system performance; scalability in dynamic environments.

Privacy-preserving frameworks that combine multiple techniques (e.g., k-anonymity, federated learning, and differential privacy) ensure that location data remains secure without compromising system performance.

Vulnerable to attacks targeting anonymization algorithms (e.g., re-identification attacks) and federated learning poisoning.

Indoor localization, mobile applications, location-based services, and data privacy in crowdsensing and IoT systems.

Location fingerprinting and anonymization¹¹⁷

Uses location fingerprinting combined with anonymization techniques to protect user privacy in vulnerable fingerprint-based systems.

Protects user privacy by anonymizing location fingerprints, preventing tracking or reidentification.

Vulnerable to attacks like Wi-Fi AP spoofing that can disrupt the fingerprinting accuracy and compromise security.

Location fingerprinting can be enhanced with anonymization techniques, such as k-anonymity, to mitigate risks of tracking or re-identification in Wi-Fi-based systems.

Spoofing attacks can mislead fingerprint matching and reduce system reliability.

Indoor navigation, Wi-Fi-based positioning systems, and secure location fingerprinting in public and private spaces.

Spoofing attack detection and prevention^95,104

Detection of spoofed tags using UWB-based systems and time-of-arrival (ToA) or time-difference-of-arrival (TDoA) methods.

High accuracy in detecting spoofed tags with sub-meter precision prevents malicious manipulation of location data.

Limited to specific technologies (e.g., UWB), real-time detection may be challenging, and scalability for large networks is difficult.

Spoofing detection systems using ToA and TDoA methods provide sub-meter localization accuracy and help mitigate the impact of spoofing attacks.

Vulnerable to advanced spoofing techniques that manipulate time-of-arrival or signal-to-noise ratios, potentially evading detection.

High-precision localization in IoT systems, security in navigation systems, anti-spoofing for UWB-based location systems, secure positioning in military or asset tracking applications.

Energy efficiency and scalability solutions^58,79,93,94

Focus on improving the energy efficiency and scalability of privacy-preserving localization systems.

Reduces energy consumption, improves system efficiency, and addresses scalability issues in dynamic environments.

Computational and communication overheads may still hinder real-time performance in large-scale, dynamic environments.

Energy-efficient techniques can significantly improve system scalability, though challenges in real-time computation and communication efficiency remain.

Attacks that drain energy resources or exploit system inefficiencies can cause service disruptions.

Energy-efficient positioning systems, smart grid applications, low-power IoT networks, and real-time localization in large-scale environments.

Privacy-preserving wireless sensing and BLE security (2025)^{127,128,129,130,131}

CSI-based sensing for human activity recognition (HAR), BLE beacon privacy protection, TEE-based privacy-preserving location queries, and MAC de-randomization for single-station user identification.

Leverages existing Wi-Fi and BLE infrastructure, enhances privacy without requiring additional hardware, enables privacy-preserving location queries with revocability.

CSI-based sensing may suffer from noise interference, BLE beacon security solutions may introduce power consumption trade-offs, and TEE-based queries require higher server-side processing costs.

Wi-Fi CSI can improve signal processing precision in HAR applications. BLE beacons require improved randomization techniques to avoid tracking risks. TEE-based solutions can securely handle location queries while maintaining revocability. Mobile single-station identification techniques reduce infrastructure requirements while improving accuracy.

Privacy concerns in CSI-based HAR, BLE beacon tracking vulnerabilities, security challenges in outsourced location queries, and MAC de-randomization risks.

Smart environments, privacy-preserving BLE-based tracking, secure location-based services, privacy-aware IoT-based indoor positioning, and non-intrusive human activity recognition.