Traditional multi-beam antennas in VHTS systems typically operate in the Ka-band and are used for extensive area coverage, employing a 7-color frequency reuse scheme that covers over 500 beams. However, current antenna technologies-namely multi-aperture and passive multi-feed multi-beam antennas-encounter substantial performance issues when tasked with covering large areas. Multi-aperture antennas, which often use three to four reflector antennas, struggle with reduced beam gain at the edges, beam deformation, and increased sidelobes. Conversely, passive multi-feed systems, which rely on a waveguide beam forming structure, become overly complex and difficult to manage when the feed count exceeds 500, a common requirement for large-scale satellite engineering.

Active multi-feed antennas, which are widely implemented in mobile communication satellites, were introduced to address these challenges. These antennas use a large mesh reflector combined with a multi-feed array to achieve high levels of sub-beam overlap, which optimizes both the gain and sidelobes of the beams through comprehensive array synthesis. However, the 12-color frequency reuse scheme used by these systems results in a Carrier-to-Interference (C/I) ratio of about 12 dB, which is insufficient for VHTS applications that require a C/I ratio of 15 dB using a 7-color frequency reuse scheme. Additionally, the complexity of the beam-forming network increases dramatically with the number of beams required.

To resolve these issues, the research team has proposed a new design method that incorporates multi-target cooperation and multi-feed amplitude and phase-weighted optimization algorithms. This approach optimizes the amplitude and phase excitation coefficients of the feeds, and establishes objective functions based on the number of feeds, the gain of the synthesized beams, and the required C/I ratio. The method also includes the use of a convolutional self-encoder surrogate model based on artificial intelligence (AI) to efficiently find the optimal beam excitation coefficients, which accelerates the optimization process significantly.

The proposed method involves an eight-step process, beginning with the GRASP model to analyze key beam parameters. The researchers then construct and train a convolutional auto-encoder model to identify and utilize nonlinear parameters in the optimization. The max-min algorithm is employed to refine these parameters iteratively, and the final optimized beam patterns are validated using the GRASP model.

When applied to VHTS technical requirements in simulations, the designed active multi-beam antenna showed impressive performance. It achieved ultra-high gain exceeding 50 dBi and a C/I ratio greater than 18 dB, all while covering nearly a thousand beams. These results suggest that this antenna design could support communication capacities at the terabits per second (Tbps) level, making it a promising solution for the next generation of satellite networks.

Research Report:An Active Multi-Beam Antenna Design Method and Its Application for the Future 6G Satellite Network

Related Links
Beijing Institute of Technology
Space Technology News - Applications and Research