GrOpt

GrOpt optimizes time-optimal gradient waveforms for magnetic resonance imaging (MRI) acquisitions.

Key Features:

• Time-Optimal Gradient Design: Generates gradient waveforms optimized for minimum time while meeting user-defined encoding and hardware constraints.
• ADMM Optimization: Solves the constrained waveform optimization problem using the alternating direction method of multipliers (ADMM).
• Constraint Support: Enforces constraints including encoding requirements, error correction mechanisms, patient comfort considerations, and peripheral nerve stimulation limits.
• Motion Compensation: Incorporates arbitrary motion (gradient moment) compensation into the waveform design.
• On-Scanner Real-Time Generation: Produces gradient waveforms directly on MRI scanner hardware and has been validated across multiple vendors.
• Performance: Produces acceptable gradient waveforms within approximately 1–40 milliseconds, with a trade-off between computation time and gradient raster time.
• Implementation and Interfaces: Implemented primarily in C with Matlab and Python wrappers.
• Versatility Across Pulse Sequences: Applicable to a wide array of MRI pulse sequences and encoding schemes.

Scientific Applications:

• On-scanner pulse sequence optimization: Enables on-the-fly synthesis and updating of gradient waveforms during MRI scans.
• High temporal resolution and complex encoding: Supports imaging techniques that require precise temporal control and advanced encoding schemes.
• Motion-robust dynamic imaging: Facilitates motion-compensated dynamic imaging through integrated gradient moment compensation.

Methodology:

GrOpt defines acquisition-specific constraints and uses ADMM to iteratively refine waveform designs to achieve time-optimal solutions that satisfy safety and performance constraints.