Designing a Controller with Image-based Pipelined Sensing and Additive Uncertainties

Abstract

Pipelined image-based control uses parallel instances of its image-processing algorithm in a pipelined fashion to improve the quality of control. A performance-oriented control design improves the controller settling time with each additional processing resource, which creates a resources-performance trade-off. In real-life applications, it is common to have a continuous-time model with additive uncertainties in one or more parameters that may affect the controller performance and the aforementioned trade-off. We present a robustness analysis framework for performance-oriented pipelined controllers with additive model uncertainties. We present a technique to obtain discrete-time uncertainties based on the continuous-time uncertainties for given uncertainty bounds. To benchmark such uncertainty bounds for a real system, we consider uncertainties in one element of the system, potentially caused by multiple uncertain parameters in the model. Robustness and its impact in the trade-off analysis are studied. We also provide a robustness-oriented pipelined controller design that takes into account the benchmarked uncertainties. Our results show that in performance-oriented designs, the tolerable uncertainties for a pipelined controller decrease when increasing the number of pipes. In robustness-oriented designs, the controller robustness is enhanced with each newly added pipe. We show the feasibility of our technique by implementing a realistic example in a Hardware-in-the-Loop simulation.

Reconfigurable Pipelined Control Systems

“…The choice of γ is made based on the requirement on sampling period h (which depends on the system dynamics ( 1)) and estimation of τ on a given platform. A proof-of-concept implementation architecture in [2] uses time-triggered activations. The cores are denoted by Core 1 to Core γ, where Core 1 is referred to as the base core.…”

“…System ( 9) is in standard discrete-time form for which standard discrete-time control design techniques such as LQR [2] can be used in the form u(kh) = Kz(kh), (10) where K is the state feedback gain to be designed. We apply u(t) = 0 for t < τ since we need to wait for the first sample to be processed before the first feedback-based actuation can take place.…”

“…Clearly, the measured signal x is six frames old and there are other older input signals in the control law. K is designed using the technique presented in [2] leading to K = −30.9 −7.3 −2.7 −1.9 , which results in a settling time of 123.5 ms.…”

“…Such sensing technologies allow to extract a rich set of information from the environment and is instrumental to develop reliable functionality in various modern systems. An example is image processing in the loop where a camera and an image processing algorithm are used for sensing [2]. The key challenge in designing such systems is to deal with the long sensing delay resulting from the heavy computation demand for sensor-data processing.…”

“…The current work focuses on how the proposed method can be implemented on embedded platforms. Our implementation is based on a proof-of-concept implementation architecture for pipelined control in [2].…”