*2.4. Potential Smart Campus LPWAN Applications*

Although a smart campus can support multiple indoor applications [70], in general, in such environments IoT nodes have access to power outlets and their communications can be usually easily handled with already common communications transceivers (e.g., Wi-Fi, Bluetooth, Ethernet). In contrast, this article focuses on the challenging environments that arise outdoors due the usual dependency on batteries to run IoT nodes and the need for exchanging data at relatively long distances (at least several hundred meters and up to 2 km), where LPWAN devices outperform other popular communications technologies.

The following are some of the most relevant outdoor applications that have already been implemented by using LPWAN technologies:


### *2.5. Smart Campus Modeling and Planning Simulators*

A smart campus is similar to an urban microenvironment where different buildings coexist with streets, open areas, parking lots, trees, benches, or people, among others. Different propagation channel models have been presented in the literature to characterize electromagnetic propagation phenomena in this type of scenarios, ranging from empirical methods to deterministic methods based on Ray Tracing (RT) or Ray-Launching (RL) approaches.

Empirical methods are based on measurements and their subsequent linear regression analysis. These methods are accurate for scenarios with the same characteristics as the real measurements, but their main disadvantages are the lack of scalability and the expensive cost in time and resources that is required to perform a measurement campaign. For example, in [97,98] empirical path loss models in a typical university campus are proposed for a frequency of 1800 MHz. The authors of [99] present a spatially consistent street-by-street path loss model for a 28 GHz channel in a micro-cell urban environment. The main drawback of these results is that the suitability of these models for path loss prediction has yet to be confirmed by the literature in other campuses.

On the other hand, deterministic methods are based on Maxwell's equations, which provide accurate propagation predictions. These approaches usually consider the three-dimensional geometry of the environment and model all propagation phenomena in the considered scenario. Their main drawback is the required computational time, which may not be afforded in large and complex scenarios. For this reason, RT or RL techniques, which are based on Geometrical Optics (GO) and the Uniform Theory of Diffraction (UTD), have been widely used for radio propagation purposes as an accurate approximation of full-wave deterministic techniques. For example, the authors of [100] modeled the dominant propagation mechanisms using advanced RT simulations in an urban microenvironment at 2.1 GHz, showing that diffuse scattering plays a key role in urban propagation. In addition, in [101] the importance of scattering when analyzing outdoor environments in the presence of trees is reported. In [102,103] the authors present propagation analyses (at 949.2 MHz and 2162.6 MHz, respectively) that make use of RT tools in a university campus with a Universal Mobile Telecommunications System (UMTS) base station placed on rooftops. Similarly, in [104], the radio-planning analysis of a Wi-Fi network in a university campus is presented using an RT approach. In addition, it has also been reported in the literature the calibration of RT simulators for millimeter-wave propagation analyses based on the measured results in a university campus at 28 GHz [105] and 38 GHz [106]. However, all the presented works are focused on micro-cellular environments and on analyzing the wireless propagation channel between a base station at a certain height in a building and a client in a pedestrian street. In this work, a radio wave propagation analysis for the connectivity of IoT LoRaWAN-based nodes by means of an in-house developed RL algorithm is reported (which is later detailed in Section 4).
