What Is Phase Estimation?

Phase estimation is the process of determining the instantaneous phase of a signal from observed data, typically in the presence of noise, frequency offsets, and other impairments. It is a core problem in signal processing and communications, arising wherever a receiver must track or recover the phase of a carrier wave to correctly demodulate data. The estimated phase is used to align the local reference at the receiver with the incoming signal, enabling coherent detection and accurate symbol decisions. Phase estimation draws from statistical estimation theory, including maximum likelihood, Bayesian, and Kalman filtering frameworks, and is implemented in both analog phase-locked loops and digital signal processing algorithms.

Phase is, by nature, a circular quantity: it wraps around at 2π, which makes estimation and tracking more analytically complex than estimating linear parameters such as amplitude. The discipline addresses both static phase offset estimation (determining a fixed unknown phase) and dynamic phase tracking (following a phase that varies over time due to oscillator drift, Doppler shift, or transmitter instability). These two regimes call for different algorithm architectures.

Maximum Likelihood Phase Estimation

Maximum likelihood (ML) phase estimation selects the phase value that maximizes the conditional probability of the observed data given that phase. In its decision-aided form, known as decision-aided maximum likelihood (DA-ML) estimation, symbol decisions from a previous step are used to remove data modulation from the received signal, leaving a residual from which the carrier phase can be estimated directly. DA-ML has become widely used in high-speed optical coherent systems operating with quadrature phase-shift keying (QPSK) and higher-order quadrature amplitude modulation (QAM) formats. Research published in IEEE/ACM Transactions on Photonics on decision-aided carrier phase estimation for coherent optical communications demonstrates that DA-ML significantly extends laser linewidth tolerance compared to simpler block-power schemes.

Phase Estimation in Coherent Optical Communications

Coherent optical communication systems rely on phase estimation to recover data from phase-modulated and amplitude-phase-modulated signals transmitted over fiber. As optical amplifiers and fiber dispersion are compensated digitally at the receiver, the residual carrier phase, contributed by laser linewidth in both the transmitter and local oscillator, must be estimated and removed on a per-symbol or per-block basis. Digital signal processing has expanded the set of phase estimation methods available beyond classical phase-locked loop approaches to include block averaging, Viterbi-Viterbi fourth-power algorithms, and multi-stage hybrid estimators. The survey Phase Estimation Methods for Optical Coherent Detection Using Digital Signal Processing on IEEE Xplore provides a systematic comparison of these approaches for M-ary PSK and QAM systems.

Statistical Foundations and Tracking Algorithms

Beyond communications, phase estimation appears in radar, sonar, power systems, and spectral analysis. In power systems, estimating the phase angles of voltage and current phasors at multiple buses is necessary for state estimation and grid synchronization. In radar, precise phase estimation enables coherent integration over multiple pulses and Doppler velocity measurement. Kalman filters and particle filters extend phase estimation to nonlinear and non-Gaussian settings, tracking phase that evolves according to a known dynamic model, such as a Wiener phase noise process. The IEEE Xplore paper on maximum likelihood estimation of Wiener phase noise variance addresses the interplay between phase noise characterization and estimator design in coherent optical receivers.

Applications

Phase estimation has applications in a wide range of technical domains, including:

• Carrier phase recovery in wireless and optical communication receivers
• Radar and sonar signal processing for coherent detection and Doppler measurement
• Phasor measurement units in power grid monitoring and state estimation
• Magnetic resonance imaging, where phase maps reveal tissue properties
• Interferometric sensing for surface profiling and gravitational wave detection