Gridless Maximum Likelihood One-Bit Direct Position Determination

doi:10.1109/LSP.2024.3491020

Gridless Maximum Likelihood One-Bit Direct Position Determination

2024-01-01

0

PRE

AI

DOI：10.1109/LSP.2024.3491020

Original

Original request for help

Save

Abstract

En 中文

Direct position determination (DPD) (a.k.a. direct localization) offers enhanced precision over traditional two-step approaches. This technique, however, involves considerable communication overhead for transmitting raw data. Low-bit direct localization methods have recently been introduced to address this issue. In this letter, we present a gridless, one-bit maximum likelihood (ML) approach for the direct localization of an orthogonal frequency division multiplexing (OFDM) signal source. A recent majorization-minimization (MM) algorithm introduced a surrogate function for the log-likelihood function, which lacks a closed-form optimal solution and requires exhaustive searches at each iteration. Our method improves upon this algorithm by developing a refined surrogate function that yields a closed-form optimal solution, thereby eliminating the need for exhaustive searches. Accordingly, the proposed MM approach can eliminate grid quantization errors (GQE) by eliminating the search process. Simulation results validate the proposed method's efficacy in mitigating GQE and its efficiency in scenarios with densely populated search grids.

Keywords:

Location awareness

OFDM

Quantization (signal)

Signal processing algorithms

Maximum likelihood estimation

Closed-form solutions

Base stations

Minimization

Upper bound

Time measurement

Direct position determination

majorization-minimization

one-bit quantization

maximum likelihood estimation

localization

Gridless Maximum Likelihood One-Bit Direct Position Determination

Abstract

Journal

Organization