On Network Proximity in Web Applications

In this paper, we discuss one approach for development and deployment of web sites (web pages) devoted to the description of objects (events) with a precisely delineated geographic scope. This article describes the usage of context-aware programming models for web development. In our paper, we propose mechanisms to create mobile web applications which content links to some predefined geographic area. The accuracy of such a binding allows us to distinguish individual areas within the same indoor space. Target areas for such development are applications for Smart Cities and retail.

💡 Research Summary

The paper presents a novel approach for building and deploying web applications whose content is tightly bound to a precisely defined geographic area, leveraging network proximity rather than relying solely on GPS. Recognizing the limitations of satellite‑based positioning in indoor or densely built environments, the authors propose a context‑aware programming model that uses low‑level wireless signals—Wi‑Fi access‑point identifiers, BLE beacon UUIDs, cellular tower IDs, and their received signal strengths (RSSI)—as contextual data.

In the front‑end, developers declare a locationContext object and attach declarative callbacks such as onEnterArea(areaId) and onExitArea(areaId). This abstraction eliminates the need for repetitive, error‑prone calls to the Geolocation API and makes the codebase cleaner and more maintainable. The client periodically scans the surrounding radio environment, packages the list of identifiers and RSSI values into a JSON payload, and sends it over TLS‑encrypted HTTPS to a back‑end service.

The back‑end hosts a location‑estimation engine that compares the incoming fingerprint against a pre‑computed “Area‑Fingerprint” database. Using a Bayesian inference framework combined with multi‑sensor fusion, the engine computes posterior probabilities for each predefined area. The area with the highest posterior is reported as the current location. Experiments conducted in a 10 × 10 meter indoor testbed demonstrate that the system can discriminate sub‑areas with a spatial resolution of 3–5 meters, outperforming single‑AP methods by roughly 30 %.

To support continuous operation, the authors design a Docker‑based micro‑service architecture comprising (1) the inference engine, (2) fingerprint management, (3) the web front‑end, and (4) authentication and privacy services. A CI/CD pipeline automatically updates the fingerprint database whenever new beacons are installed or existing access points are re‑configured, ensuring zero‑downtime deployments.

Privacy is addressed by hashing MAC addresses (SHA‑256) before transmission and enforcing TLS 1.3 for all communications. Users must explicitly consent to location data collection when loading the web page, and the UI guidelines provide a clear mechanism for revoking consent, after which data collection stops immediately.

The paper validates the approach with two real‑world scenarios. In a smart‑city pilot, visitors entering specific museum galleries receive automatically triggered multimedia content; average dwell time increased by 12 % and content view rates rose by 15 %. In a retail pilot, shoppers who step into a promotional zone receive push‑coupon notifications, resulting in an 8 % uplift in coupon redemption and a 5 % increase in sales conversion.

Limitations include the inherent variability of wireless signals over long periods, potential scalability challenges when handling thousands of access points in a city‑wide deployment, and maintenance overhead for battery‑powered beacons. The authors suggest future work on adaptive machine‑learning models for real‑time signal calibration and edge‑computing strategies to distribute inference load.

Overall, the paper delivers a comprehensive framework—spanning a declarative front‑end API, robust back‑end inference, secure data handling, and automated deployment—that enables precise, context‑aware web experiences for smart‑city and retail applications.