LyceanEM: A python package for virtual prototyping of antenna arrays, time and frequency domain channel modelling

The design of antenna arrays to meet complex requirements in sensors and communications depends upon access to robust tools to access and prototype potential antenna arrays on their intended structures without the need for in depth antenna design. Together with the simulation of wider communications and imaging problems, there is a pressing need for efficient electromagnetics tools capable of scaling from individual antennas to ultra large scale antenna arrays with 1000s of individual elements.


Statement of need
LyceanEM is a Python library for modelling electromagnetic propagation for sensors and communications. Frequency Domain and Time Domain models are included that allow the user to model a wide array of complex problems from antenna array architecture and assess beamforming algorithm performance to channel modelling. The model is built upon a ray tracing approach, allowing for efficient modelling of large, low density spaces.
LyceanEM relies upon the Numba package to provide CUDA acceleration of electromagnetics, calculating antenna and antenna array patterns, scattering and aperture projections. This has been used in a number of scientific publications (Pelham, 2022;Pelham, Freire, et al., 2021; and has been used in a tutorial on Antenna Array Design for Complex Platforms at Radar 2022. This capability in an open source package enables exciting research by academics and professional engineers alike.
LyceanEM is also being used for ongoing multidisciplinary research combining channel modelling and spatial mapping using computer vision. The flexible and efficient nature of the scattering model allows for exciting exploration of the signal sources in the local environment on low power computing devices.
The benefit of this emphasis on rapid virtual prototyping is to allow the user to quickly establish the potential performance for a desired aperture and frequency, on a desired platform with relatively little design effort. Comparatively, the otherwise excellent commercial solvers like CST, HFSS, FEKO etc can provide excellent simulation fidelity, but require a significant design investment before the simulation can be run. This lack of coverage leads to an uncertain design process for antenna arrays in which the requirements for an antenna array can be specified without reference to the physical limitations imposed by the desired aperture size, location, polarisation, beamforming envelope etc. LyceanEM allows these factors to all be predicted rapidly, and the beamforming architecture to be simulated in a realistic way, providing crucial design insight at a low cost.

Usage Examples
While some usage examples are presented here, many more are included in the documentation for LyceanEM, which can be found at https://documentation.lyceanem.com/en/latest/.

Virtual Prototyping and Antenna Array Beamforming Research
The initial intended use case for LyceanEM was virtual prototyping for antenna arrays and apertures for sensors and communications. As show in Figure 1 & Figure 2. This allows for antenna array patterns to be predicted extremely quickly compared to the time required for antenna design, and simulation on the platform of interest. This enables research into novel conformal antenna array configurations, and modelling of the performance of radar antenna arrays for autonomous vehicles research. This is the only package offering this capability from such a limited information set. This process allows the researcher or engineer to assess the maximum achievable beamforming envelope for the aperture ( Figure 3), then predict the antenna array pattern with functions supporting beamsteering to points of interest to generate an accurate prediction of the beamformed coverage ( Figure 4). This has been demonstrated in published research for both antenna arrays (Pelham, Freire, et al., 2021), and for reflector antennas (Pelham, 2022). The antenna array simulation allows the researcher to define the array with any combination of polarisation and excitation functions, providing a powerful tool for antenna array research.    In addition to modelling of antenna arrays with conventional polarisation and propagation modes, LyceanEM supports the modelling and analysis of novel propagation models such as Orbital Angular Momentum states (Allen et al., 2019) for communications links, and allowed the prediction of the supportable OAM modes of an aperture, and the fourier analysis of the modal spectrum produced both by LyceanEM and the measured antenna patterns.

Frequency & Time Domain Channel Modelling
LyceanEM can also be used as a more general electromagnetic model, allowing the definition and simulation of complex channel models. In a published example, the Frequency domain model predicted the scattering parameters produced when illuminating a rotating metal plate with a horn antenna with a root mean square (RMS) error of -69dB between the predicted scattering parameters and the measured data. . This setup is shown in Figure 5 with the scattering plate at an angle of 45 degrees, and the transmitting and receiving horn antennas shown. The resultant scattering with variation of normalised scattering angle (0 degrees when plate is 45 degrees offset from both transmitter and receiver) shows the comparison between the measured scattering at 26GHz, and that predicted by LyceanEM in Figure 6.