As result of technology scaling, single-chip multicore power density increases and its spatial and temporal workload variation leads to temperature hot-spots, which may cause nonuniform ageing and accelerated chip failure. These critical issues can be tackled by closed-loop thermal and reliability management policies. Model predictive controllers (MPC) outperform classic feedback controllers since they are capable of minimizing performance loss while enforcing safe working temperature. Unfortunately, MPC controllers rely on a priori knowledge of thermal models and their complexity exponentially grows with the number of controlled cores. In this paper, we present a scalable, fully distributed, energy-aware thermal management solution for single-chip multicore platforms. The model-predictive controller complexity is drastically reduced by splitting it in a set of simpler interacting controllers, each one allocated to a core in the system. Locally, each node selects the optimal frequency to meet temperature constraints while minimizing the performance penalty and system energy. Comparable performance with state-of-the-art MPC controllers is achieved by letting controllers exchange a limited amount of information at runtime on a neighborhood basis. In addition, we address model uncertainty by supporting learning of the thermal model with a novel distributed self-calibration approach that matches well the controller architecture.

A novel speed sensorless indirect field-oriented control for the full-order model of the induction motor is presented. It provides local exponential tracking of smooth speed and flux amplitude reference signals together with local exponential field orientation, on the basis of stator current measurements only and under assumption of unknown constant load torque. Speed estimation is performed through a reduced-order adaptive observer based on the torque current dynamics, while no flux estimate is required for both observation and control purposes. The absence of the flux model in the proposed algorithm allows for simple and effective time-scale separation between the speed-flux tracking error dynamics (slow subsystem) and the estimation error dynamics (fast subsystem). This property is exploited to obtain a high performance sensorless controller, with features similar to those of standard field-oriented induction motor drives. Moreover, time-scale separation and physically-based decomposition into speed and flux subsystems allow for a simple and constructive tuning procedure. The theoretical analysis based on the singular perturbation method enlightens that a persistency of excitation condition is necessary for the asymptotic stability. From a practical viewpoint, it is related to the well-known observability and instability issues due to a lack of back-emf signal at zero-frequency excitation. A flux reference selection strategy has been developed to guarantee Persistency of excitation in every operating condition. Extensive simulation and experimental tests confirm the effectiveness of the proposed approach.

The throttle control plays an important role in the development of drive by wire (DBW) systems. The position control of the throttle valve is quite a complex problem, due to application constraints and system characteristics. A cascaded control structure including a nonlinear trajectory generator filter is adopted, allowing each different control problem to be solved with the most suitable control algorithm and implementation technology. In this regard, the use of variable structure control techniques is the key element to reaching the solution. Extensive simulation tests are reported to show the performance of the proposed control algorithm. A prototype controller is presented. The experimental implementation of the controller for a step from 2.5/spl deg/ to 85.5/spl deg/ indicates a very smooth position trajectory with a maximum dynamic position error of 7/spl deg/. Application specifications are fully satisfied both in terms of control performance and controller cost.