As Smart Cities become increasingly connected, thousands of IoT devices are being deployed in streets, buildings, vehicles, and public infrastructure. These devices collect valuable information about traffic, energy, mobility, environmental conditions, and urban services. But collecting data is only part of the challenge. In many applications, it is equally important to know where the device, vehicle, or object is located.
This is where IoT localization comes into play.
IoT localization refers to the technologies and algorithms used to determine the position of connected devices. Depending on the application, localization can rely on signal strength, signal arrival time, network connectivity, satellite systems, or artificial intelligence.
Today, these approaches can be broadly divided into range-based, range-free, hybrid, and AI-ML-based localization.
1. Range-Based Localization: Measuring Distance or Direction
Range-based techniques estimate the distance or angle between devices using physical characteristics of wireless signals.
The most common approaches include:
*RSSI – Received Signal Strength Indicator
*ToA – Time of Arrival
*TDoA – Time Difference of Arrival
*AoA – Angle of Arrival
* Received Signal Strength Indicator (RSSI): RSSI-based localization estimates the distance between a transmitter and a receiver from the strength of the received wireless signal.
Its major advantage is simplicity. RSSI measurements are already available in many wireless communication systems, so additional specialized hardware may not be required.
However, RSSI can be strongly affected by the surrounding environment. Buildings, obstacles, interference, fading, and multipath propagation can change signal strength and reduce positioning accuracy.
This is particularly important in dense urban environments, where buildings and moving objects can significantly influence wireless signals.
* Time of Arrival (ToA) : ToA estimates distance by measuring how long a signal takes to travel between a transmitter and receiver.
* Time Difference of Arrival (TDoA): TDoA compares the arrival times of a signal at different receivers.
These techniques can provide high accuracy, but they generally require precise timing and synchronization and may require specialized hardware.
* Angle of Arrival (AoA): Knowing the Direction: AoA determines the direction from which a wireless signal arrives.
By combining measurements from several reference points, the system can estimate the position of a device. AoA can provide accurate positioning, but antenna arrays or other specialized equipment may increase the complexity and cost of deployment.
2. Range-Free Localization: Simpler and Less Expensive
Unlike range-based methods, range-free localization does not directly measure distance or angle.
Instead, it uses information such as:
* Network connectivity
* Communication hops
* Anchor-node positions
* Geometric relationships between nodes
Some well-known range-free algorithms are DV-Hop, Centroid, SL-Free and APIT.
* DV-Hop
DV-Hop estimates the position of an unknown node by analyzing the number of communication hops separating it from known anchor nodes.
* Centroid
The Centroid method estimates a node’s position using the coordinates of neighboring anchor nodes.
* APIT
Approximate Point-In-Triangulation (APIT) determines whether a node is located inside or outside triangles formed by anchor nodes.
* SL-Free
Self-Localisation Free algorithm exploits network connectivity and the known positions of anchor nodes to estimate the positions of unknown sensors. Its objective is to obtain accurate position estimates while maintaining low resource requirements.
Range-free approaches are attractive because they generally require less specialized hardware and lower resources. However, their positioning accuracy is usually lower than that of range-based techniques.
In simple terms:
Range-based → higher potential accuracy, greater hardware requirements Range-free → simpler deployment, lower cost, but generally lower accuracy
3. Hybrid Localization: Combining the Best of Several Technologies
What happens when one technology is not reliable enough?
A solution is to combine multiple localization technologies.
Hybrid localization can combine information from /GNSS, /RSSI, /Wi-Fi, /Cellular networks, /BLE, and /LPWAN.
Other wireless measurements; For example, Global Navigation Satellite System (GNSS); may provide accurate outdoor positioning, while Wi-Fi or Cellular positioning can help when satellite signals are blocked by buildings.
This is particularly useful in Smart Cities because urban environments are highly variable.
A localization system may therefore adapt its strategy according to:
* The environment
* Required accuracy
* Available infrastructure
* Device capabilities
* Energy consumption The result is a more flexible and robust positioning system.
4. AI and Machine Learning for IoT Localization
Another important development is the use of Artificial Intelligence (AI) and Machine Learning (ML).
Traditional localization techniques often depend on mathematical models describing how wireless signals propagate. However, real urban environments are complicated. Buildings, vehicles, obstacles, interference, and multipath effects can make these models difficult to apply accurately.
Machine learning offers another possibility: instead of relying entirely on predefined models, algorithms can learn the relationship between signal measurements and geographical positions from collected data.
ML techniques have been explored for:
* RSSI-based localization
* Wi-Fi fingerprinting
* Sensor fusion
* Location prediction
* IoT positioning
This makes AI-based localization particularly promising for Smart Cities, where large amounts of heterogeneous sensor data are continuously generated. The long-term goal is to develop localization systems that can learn, adapt, and improve their performance according to changing urban conditions.
5. LPWAN and GNSS: Large-Scale and Energy-Efficient Positioning
Smart Cities may contain thousands or even millions of connected devices. Many of these devices need to operate for long periods using limited battery power.
This has increased interest in Low-Power Wide-Area Network (LPWAN) technologies.
Technologies such as LoRaWAN are designed to provide long-range communication while keeping energy consumption relatively low.
Their signals can also be exploited for positioning, providing an alternative or complement to conventional satellite-based positioning.
* Global Navigation Satellite Systems (GNSS)
Global Navigation Satellite Systems (GNSS), including GPS and Galileo, remain fundamental for outdoor positioning.
However, GNSS can face difficulties in dense urban areas because buildings can block or reflect satellite signals. These effects can lead to signal attenuation, multipath propagation, and reduced positioning reliability.
GNSS can also be relatively energy-demanding for battery-powered IoT devices. Combining GNSS with LPWAN-based positioning can therefore provide an interesting balance between coverage, energy consumption, infrastructure requirements, and positioning accuracy.
6. Emerging Technologies: UWB and LEO-PNT
IoT localization is also moving toward new technologies capable of improving accuracy and resilience.
* Ultra-Wideband (UWB)
Ultra-Wideband (UWB) has attracted significant attention for high-precision positioning.
It is particularly suitable for indoor and short-range applications where very accurate location information is required.
Potential applications include:
* Indoor navigation
* Asset tracking
* Smart buildings
* Robotics
* Industrial IoT
The main challenge is that UWB deployments may require dedicated anchor infrastructure, which can increase installation costs.
* LEO-PNT
Another emerging direction is Low Earth Orbit Positioning, Navigation and Timing (LEO-PNT).
LEO-PNT uses signals from satellites in Low Earth Orbit to support positioning, navigation, and timing.
Researchers are investigating LEO-PNT as a potential complement to GNSS. Its different satellite characteristics and geometry could provide additional positioning capabilities, particularly in challenging environments. Although promising, LEO-PNT is still an emerging research area, with challenges related to infrastructure, receivers, signal design, coverage, and integration with existing GNSS systems.
7. Which Localization Technology Is Suitable for a Smart City?
Different urban applications require different levels of positioning accuracy and reliability.
GNSS: Best suited for outdoor positioning and applications requiring global coverage.
Wi-Fi: Useful for indoor positioning and fingerprint-based localization, particularly where Wi-Fi infrastructure already exists.
BLE: Suitable for proximity and indoor positioning, especially with Bluetooth beacons.
ZigBee: Interesting for low-power wireless sensor networks.
Cellular Networks: LTE and 5G can support positioning while taking advantage of existing communication infrastructure.
LPWAN: Useful for large-scale, low-power IoT deployments.
UWB: Particularly attractive when high positioning accuracy is required over relatively short distances.AI/ML: Can enhance several of these technologies by learning from historical measurements and adapting to complex environments.
8. Choosing the Right Localization Approach
There is no single localization technology that is ideal for every Smart City application.
The appropriate solution depends on several factors:
* For accuracy: How precise must the location be?
* For coverage: How large is the area?
* For energy: Is the device battery-powered?
* For Cost: How much infrastructure can be deployed?
* For environment: Indoor, outdoor, or dense urban area?
* For scalability: Can the system support thousands of devices?
* For robustness: Can it operate when signals are blocked or degraded?
* For infrastructure: What positioning technologies are already available?
For example, a smart parking system may require relatively precise positioning, while an environmental sensor may only need to identify the general area where it is located.
9. The Future: Toward Intelligent and Multi-Source Localization
The evolution of IoT localization shows a clear trend: moving from single-technology positioning toward intelligent, multi-source systems.
GNSS remains essential for outdoor positioning. Wireless technologies such as Wi-Fi, BLE, cellular networks, and LPWAN provide complementary solutions. UWB offers high-precision positioning for specific applications, while LEO-PNT represents an emerging alternative or complement to GNSS.
At the same time, AI and machine learning are changing the way localization systems process and interpret positioning data.
The future of Smart City localization will therefore likely involve systems capable of combining several sources of information and dynamically selecting the most appropriate positioning strategy.
Rather than asking: Which localization technology is the best?
the more relevant question is:
Which combination of technologies can provide the most reliable and efficient localization for a specific Smart City application? This shift toward hybrid, adaptive, and intelligent localization could play an important role in building more connected, efficient, and resilient cities.