Simulation and Performance Analysis of Time-Modulated Array Beamforming

Phased-array beamforming steers a beam electronically toward a Low Earth Orbit (LEO) satellite, but the phase shifter needed at every element makes the hardware costly, power-hungry, and difficult to scale for a compact Non-Terrestrial Network (NTN) user terminal. A conventional binary Time-Modulated Array (TMA) removes the phase shifter by switching each element ON/OFF in time, but its coarse two-state control caps achievable sidelobe level (SLL) and wastes radiated power in unwanted harmonics. This work develops a tri-state (+1, 0, −1) TMA for 5G New Radio (NR) NTN downlink terminals, in which each element is driven by an independent positive and negative pulse within a shared switching period. Closed-form harmonic coefficients are derived for the tri-state switching function and combined with the array steering vector to synthesise the firstharmonic radiation pattern. A Quadratic Programming Optimization (QPO) routine, warm-started with Chebyshev tapering, jointly minimises sidelobe level and half-power beamwidth (HPBW), while an Alternating Direction Method of Multipliers (ADMM) framework, realised as a Bartlett spatial-matched-filter coarse-to-fine search, adaptively tracks the satellite's direction of arrival (DoA). Simulated on a multi-element ULA driven by a realistic 5G NR OFDM waveform, the developed design reduces the sidelobe level from approximately −13 dB to −26 dB while improving output SINR and maintaining accurate beam tracking, with a beam pattern nearly indistinguishable from the ideal-DoA case, all at a considerably lower relative hardware cost than a fully phase-shifter-based array.