Past and Current Research

Controls

• Safety verification of cyber-physical systems is imperative because of the ever-increasing proliferation of smart networks. Deductive safety verification techniques typically involve some continuous analog of induction, for example, differential induction or barrier certificates. Moreover, control systems are often designed under the assumption that controllers run continuously, while the actual implementation involves a digital controller that operates at a given sample rate. Our current efforts focus on developing numerically viable safety verification techniques for geofence avoidance of unmanned aerial vehicles. Our proposed safety verification technique allows a trade-off between the size of the sample period and the convergence rate. An expression for the control law and an explicit relationship between sample period and control parameters are presented. It is shown that the proposed control law drives all state trajectories initiated in the safe set to the origin without violating safety criteria as long as the sample period remains sufficiently small. Moreover, if the control signal is bounded between the proposed limits, it is shown analytically that a pilot-based configuration also remains in the safe set [1].


• Component swapping modularity is a recently proposed concept in control networks to guarantee desired closed-loop performance using low-order modular controllers for smart components. Conventionally, complete control redesign is inevitable when swapping a system's component with a dynamically different counterpart. However, our proposed algorithm achieves desired closed-loop performance only by tuning the low-order controller of the swapped component. The major part of the control remains unchanged for all the variants of the swappable component. The proposed algorithm dramatically simplifies control design and reduces calibration time and effort, particularly, in applications from automotive industry and power networks [2, 3].


• Multivariable Newton-based extremum seeking (ES) extends the conventional gradient-based ES designs, so uniform transients are achieved for all channels of a multivariable optimization problem. The proposed design governs the system to its optimal point on a straight path, so control effort is also minimized. The Newton-based ES is robust under external disturbances or model uncertainties. Moreover, ES algorithms in general, and specifically the Newton-based design, are real-time continuous and do not require processing units which further simplifies algorithm implementation and reduces manufacturing and maintenance cost [10].

Energy Systems

• Aggregate modeling and control of thermostatically controlled loads (TCLs) provide a promising avenue to improve energy efficiency and the stability margin of power networks. We proposed a novel analytical model for a large population of heterogeneous heating/cooling units. As shown in Fig. 1(c), averaged accumulated error of the model is less than 0.3%. Simple and yet effective power control algorithms were designed using the proposed model. Our model is beneficial to utility managers as well as the end-user to increase power efficiency and reduce energy cost [5]. We designed and installed a wireless sensor network, illustrated in Fig. 1(a), to verify the effectiveness of our proposed model, experimentally.



• Power optimization of photovoltaic (PV) energy generators is essential to exploit PV power resources fully. The proposed algorithm is non-model-based, and hence can be applied to different PV systems with minimal redesign. It offers the advantages of fast convergence and guaranteed stability over a wide range of environmental conditions, and yet is simple and cost-effective to implement. Thus, power efficiency is increased, and energy cost is reduced [6, 7, 9].



• Nonlinear control and power optimization of wind energy conversion systems (WECS) guarantee maximum feasible wind power extraction under fast changing wind speeds. We use a nonlinear controller, based on the field-oriented control concept and feedback linearization, to achieve instantaneous wind power tracking. Extremum seeking maintains wind turbine energy yield at its feasible peak and also improves system robustness versus unknown disturbances. Thus, higher power efficiency is achieved, and ultimately energy cost is reduced [8, 9].



Precision Mechatronics

• High-precision contour error estimation and contouring algorithms improve the accuracy of machine tools, defined using unmeasurable contour error which equals the shortest distance from the actual position to the reference contour. Hence, we proposed a precise dynamic contour error estimate (CEE) and a novel integrated contouring algorithm to govern the contour error of a wide range of reference feedrates to zero. Our experiments showed as much as 50% reduction in contour error for highly curved fast references [4].



• Robust nonlinear control of flexible-link robots comprises a modified sliding mode control using global stabilization and nonlinear optimization. The proposed algorithm constructed based on the sliding mode control because the flexible-link nominal model has order uncertainty and is sensitive to external load disturbances. The research was conducted to exploit benefits of high-order switching surfaces to improve closed-loop system performance of a sliding mode control designed based on a low order nominal model of the flexible-link [11, 12].

References

1. A. Ghaffari, I. Abel, D. Ricketts, S. Lerner, and M. Krstić, “Safety verification using barrier certificates with application to second order systems with constrained discrete inputs,” American Control Conference, submitted.
2. A. Ghaffari and A. G. Ulsoy, “Component swapping modularity for distributed precision contouring,” IEEE/ASME Transactions on Mechatronics, to appear.
3. A. Ghaffari and A. G. Ulsoy, “LMI-based design of distributed controllers to achieve component swapping modularity,” IEEE Transactions on Control Systems Technology, to appear.
4. A. Ghaffari and A. G. Ulsoy, “Dynamic contour error estimation and feedback modification for high-precision contouring,” IEEE/ASME Transactions on Mechatronics, DOI 10.1109/TMECH.2015.2494518.
5. A. Ghaffari, S. Moura, and M. Krstić, “Modeling, control, and stability analysis of heterogeneous thermostatically controlled load populations using partial differential equations,” ASME Journal of Dynamic Systems, Measurement, and Control, vol. 137, pp. 101009(1-9), 2015.
6. A. Ghaffari, S. Seshagiri, and M. Krstić, “Multivariable maximum power point tracking for photovoltaic micro-converters using extremum seeking,” Control Engineering Practice, vol. 35, pp. 83-91, 2015.
7. A. Ghaffari, M. Krstić, and S. Seshagiri, “Power optimization for photovoltaic micro-converters using multivariable Newton-based extremum seeking,” IEEE Transactions on Control Systems Technology, vol. 22, pp. 2141-2149, 2014.
8. A. Ghaffari, M. Krstić, and S. Seshagiri, “Power optimization and control in wind energy conversion systems using extremum seeking,” IEEE Transactions on Control Systems Technology, vol. 22, pp. 1684-1695, 2014.
9. A. Ghaffari, M. Krstić, and S. Seshagiri, “Extremum seeking for wind and solar energy applications,” Mechanical Engineering, vol. 136, no. 3, pp. S13-S21, March 2014.
10. A. Ghaffari, M. Krstić, and D. Nešić, “Multivariable Newton-based extremum seeking,” Automatica, vol. 48, pp. 1759-1767, 2012.
11. A. Ghaffari and M. J. Yazdanpanah, “Computing optimized nonlinear sliding surfaces,” in Proc. of Chinese Control and Decision Conference, 2008.
12. M. J. Yazdanpanah and A. Ghafari, “Modification of sliding mode controller by using a neural network with application to a flexible-link,” in Proc. of European Control Conference, 2003.