Coherent Radar Backscatter Modelling for Drones in the OFDM Framework

Abstract

The problem statement of this thesis is to construct a simple, versatile, closed-form simulation model which can be used to model Orthogonal Frequency Division Multiplexing (OFDM) pulsed-radar reflections from drones. General simulation methodologies exist which allow a simulation designer to provide an arbitrary transmit waveform, system parameters and a set of point-scatterer positions and velocities representing complex radar targets for each time sample considered in the simulation and to generate signal reflection predictions. On the other hand, certain over-simplified methodologies exist to generate reflection predictions from propellers or drones and significantly simplified to allow a closed-form solution. The manner in which to simulate signal reflections from complex targets explored in this thesis provides a middle-ground between the over-generalized and over-simplified simulation methodologies and is adapted to the specific use-case of OFDM pulsed-radar to model drone targets. The benefit of this simulation methodology is its ease-of-use and flexibility. The solution allows one to provide system parameters, drone parameters and a drone trajectory to produce accurate backscatter predictions without having to model all the individual point-scatterers for the radar target at each point in time. Additionally, the simulation methodology presented in this thesis is derived in a separate manner than the current standard for complex target modeling which is having well-placed point-scatterers anchored to the simulated target. If one generates prediction results from two separate simulation algorithms, then their certainty in simulation accuracy increases. This thesis presents how the simulation model derived within produces almost identical results to the state-of-the-art simulation methodology, where the same complex radar target is modeled as a set of moving point-scatters with varying Radar Cross-Sections (RCS) at each point in discrete-time. Since the model derived within require significantly less setup and modeling efforts from the user and provides roughly the same computational complexity for modeling drone targets in the OFDM pulsed-radar framework, it is the better option. The main metrics used for comparison between the different simulation methodologies are the Fourier transforms of the received signal, micro-Doppler signatures and the range-speed response for the radar system. All three metrics are shown to match the state-of-the-art in the proposed simulation method.