Wireless Localization | Wireless Systems Lab

1. Ultra-Wide Band Localization

Among indoor positioning technologies that have emerged in recent years, positioning using Ultra Wide Band (UWB) is a new technology that performs better than other technologies in indoor positioning. The high data rate of UWB can reach 100Mbps, making it a suitable solution for short-distance data transmission. In addition, the high bandwidth and extremely short pulse waveforms help reduce the effects of multipath interference and facilitate TOA measurements, making UWB a desirable solution for indoor positioning.

Proposed UWB-based tracking system uses swarm of mini-quadcopters (Crazyflie 2.1) to localize and track the moving vehicle. Quadcopters wirelessly measure distance to target via UWB communications and localize the target using multi-lateration. With the localized position of target being centered, quadcopters form regular polygon shape formation. The small size of the quadcopters are safe to humans and suitable for indoor flight.

Video 1. UWB-based Multiple UAV Control System for Indoor Ground Vehicle Tracking

2. Fingerprinting Technique

Fingerprinting is one of the most exploited techniques in indoor positioning method. The wireless device collect the received signal strength indicator (RSSI) from multiple access points (AP), which represents the fingerprint of location. Server store the RSSI data at each reference point (RP) to learn the RSSI map information and then infer the device’s location.

We are studying on machine learning (ML)-based fingerprinting in a manner of privacy preserving multi-user crowd-sourcing method. Federated learning (FL) address the privacy issue in cooperation among a massive users for localization and location data processing.In FL, the edge devices update local ML model and the central server aggregates the local models. Without transmitting any raw data that may harm privacy, the server is enable to learn personal data’s characteristics.

3. Range-free Localization

Range-free localization algorithms for wireless sensor networks have been introduced with emphasis of cost- and energy-effectiveness. Range-free localization algorithms presume that the length of the shortest path for packet delivery from one to another corresponds to their Euclidean distance. The distance from a normal node to an anchor at known location is then estimated as the product of the hop count between them and the average hop progress, which is in general computed as the ratio of the distances and the hop counts between anchors.

However, anisotropic networks are frequently observed in practice due to various anisotropic factors including non-uniform node deployments, irregular radio patterns, and irregular-shaped regions. To deal with these problem, we propose a range-free localization algorithm with reliable anchor pair selection (RAPS). In the RAPS algorithm, each normal node selects reliable anchor pairs based on the average hop progresses.

Sensor fusion is achieved through a combination of different radio signals. The combination of wireless signals enables positioning while compensating for the disadvantages of each wireless signal compared to when wireless signals are used alone, enabling positioning with high accuracy in a wider range.

IMU (Inertial Measurement Units) sensors are mounted on an object to measure the direction, speed, and acceleration of the object, and are sensors that can position the object without being affected by the outside. Positioning using the IMU sensor alone can position an object accurately in a short time, but there are disadvantages in that the direction is deflected during long-time positioning, and errors occur and accumulate in the velocity and acceleration measurement values, resulting in a large positioning error. Therefore, by fusion of UWB and IMU sensors, the influence of the environment such as NLOS can be reduced and the error accumulation problem can be solved, so that the positioning performance can be improved compared to when the sensor is used alone.

Sensor fusion is achieved through a combination of different radio signals. The combination of wireless signals enables positioning while compensating for the disadvantages of each wireless signal compared to when wireless signals are used alone, enabling positioning with high accuracy in a wider range.

IMU (Inertial Measurement Units) sensors are mounted on an object to measure the direction, speed, and acceleration of the object, and are sensors that can position the object without being affected by the outside. Positioning using the IMU sensor alone can position an object accurately in a short time, but there are disadvantages in that the direction is deflected during long-time positioning, and errors occur and accumulate in the velocity and acceleration measurement values, resulting in a large positioning error. Therefore, by fusion of UWB and IMU sensors, the influence of the environment such as NLOS can be reduced and the error accumulation problem can be solved, so that the positioning performance can be improved compared to when the sensor is used alone.

4. Sensor Fusion

5. Cooperative Localization

5G new radio will provide a new paradigm in high accurate vehicle localization, reinforced by the use of large antenna arrays along with carefully designed broadband radio technology. The assessment for V2V and V2I positioning technology is still under development for future applications. Thus, a new paradigm for efficient localization in the vehicular networks is urged. Large bandwidths and large antenna arrays respectively afford the support for highly resolvable distance and angle estimation. Furthermore, seamless connectivity empowered by D2D cooperation also becomes available via the line-of-sight (LOS) radio propagation in highly dense networks. However, a high computational load still remains unravelled in the context of cooperative localization, albeit with its advantages of high-precision localization. A Deep neural network (DNN) assisted, and an alternating direction method of multipliers (ADMM) formalism for cooperative localization readily enables to recast a formulation of the decentralized optimization. They are efficient solutions of the optimization is developed in a distributed manner.

J. Eom, H. Kim, S. H. Lee, and S. Kim, “DNN-Assisted Cooperative Localization in Vehicular Networks,” Energies, Special Issue “Wireless Communication Systems for Localization,” vol. 12, no. 14, pp. 2758, Jul. 2019.
H. Kim, S. H. Lee, and S. Kim, "Cooperative localization with Distributed ADMM Over 5G-based VANETs," in Proc. 2018 IEEE Wireless Communications and Networking Conference (WCNC), Barcelona, Spain, 2018, pp. 1-5.
H. Kim, S. H. Lee, and S. Kim, “Cooperative Localization with Constraint Satisfaction Problem in 5G Vehicular Networks,” IEEE Trans. Intell. Transp. Syst., Nov. 2020.

Video 2. ADMM based vehicle cooperative positioning POC using V2X test bed