Phase Compensation Architectures for Asynchronous Decimation Delays in Multirate Feedback Loops
Dynamic Farrow filters and state observers restore feedback phase margin by compensating time-varying asynchronous decimation group delays in real time.
Multirate Feedback Loops
Control Architecture
Discrete time systems employ different sampling frequencies for various sensors and actuators to optimize processing resources. Within such a framework, multirate feedback loops allow a fast inner loop to handle high frequency dynamics while a slower outer loop manages global setpoints. This approach reduces the computational burden on the central processing unit without sacrificing stability.
This architecture is particularly effective in high precision motion control.
Stability Analysis
Aliasing and hidden oscillations represent the primary risks when mixing disparate clock domains. Stability for multirate feedback loops is verified using lifting techniques that convert the time varying system into a time invariant model. Lifting techniques confirm that the inter sample behavior remains bounded.
Data Synchronization
Zero order hold circuits and decimation filters manage the transfer of information between the different rate groups. If a sensor updates at one kilohertz and the controller runs at one hundred hertz, the multirate feedback loops must incorporate an anti aliasing filter to prevent noise folding. Jitter in the sampling clock can degrade the phase margin.
Computational Efficiency
Allocation of tasks depends on the physical time constant of the hardware being controlled. A temperature regulator does not require the same update frequency as a motor drive, making multirate feedback loops ideal for complex mechatronic assemblies.