Real-time Autonomy Research Articles

Two-stage cooperative real-time autonomy control strategy considering vulnerability of active distribution networks

A two-stage cooperative real-time autonomy control strategy considering vulnerability of active distribution networks (ADNs) is proposed to solve the problems of intensified voltage fluctuation, increased network loss, and risk of voltage violation. A vulnerability index of the power grid is proposed to optimize the model from “point” to “area”, which can more comprehensively evaluate the power quality of ADNs. The day-ahead stage, considering the action times of the on-load tap changer (OLTC) and capacitor bank (CB) and the reactive power response ability of the photovoltaic (PV) inverter, a relaxation-clustering-correction decoupling strategy based on Ward clustering is proposed to make the optimization results more reasonable; based on the curve-increasing particle swarm optimization (CIPSO) algorithm, a high-dimensional, strongly coupled mixed integer nonlinear model was solved. The real-time stage, taking the optimization results of the day-ahead stage as the benchmark data, based on the relationship between the reactive power increment and real-time output prediction error of PV inverter, a real-time autonomy control strategy that only relies on local real-time information is proposed by fusing voltage sensitivity information. The rationality and superiority of the proposed strategy are verified through the simulation analysis of the modified IEEE 33 bus system.

Fingerprints of Teaching Interactions: Capturing and Quantifying How Supervisor Regulate Autonomy of Residents in the Operating Room

Supervisors and residents agree that entrusted autonomy is central to learning in the Operating Room (OR), but supervisors and residents hold different opinions about entrustment: residents regularly experience that they receive insufficient autonomy while supervisors feel their guiding is not appreciated as teaching. These opinions are commonly grounded on general experiences and perceptions, instead of real-time supervisors' regulatory behaviors as procedures unfold. To close that gap, we captured and analyzed when and to what level supervisors award or restrain autonomy during procedures. Furthermore, we constructed fingerprints, an instrument to visualize entrustment of autonomy by supervisors in the OR that allows us to reflect on regulation of autonomy and discuss teaching interactions. All interactions between supervisors and residents were captured by video and transcribed. Subsequently a multistage analysis was performed: (1) the procedure was broken down into 10 steps, (2) for each step, type and frequency of strategies by supervisors to regulate autonomy were scored, (3) the scores for each step were plotted into fingerprints, and (4) fingerprints were analyzed and compared. University Medical Centre Groningen (the Netherlands). Six different supervisor-resident dyads. No fingerprint was alike: timing, frequency, and type of strategy that supervisors used to regulate autonomy varied within and between procedures. Comparing fingerprints revealed that supervisors B and D displayed more overall control over their program-year 5 residents than supervisors C and E over their program-year 4 residents. Furthermore, each supervisor restrained autonomy during steps 4 to 6 but with different intensities. Fingerprints show a high definition view on the unique dynamics of real-time autonomy regulation in the OR. One fingerprint functions as a snapshot and serves a purpose in one-off teaching and learning. Multiple snapshots of one resident quantify autonomy development over time, while multiple snapshots of supervisors may capture best teaching practices to feed train-the-trainer programs.

Human Decision Making Model for Autonomic Cyber Systems

Real-time autonomy is a key element for system which closes the loop between observation, interpretation, planning, and action, commonly found in UxV, robotics, smart vehicle technologies, automated industrial machineries, and autonomic computing. Real-time autonomic cyber system requires timely and accurate decision making and adaptive planning. Autonomic decision making understands its own state and the perceived state of its environment. It is capable of anticipating changes and future states and projecting the effects of actions into future states. Understanding of current state and the knowledge/model of the world are needed for extrapolating actions and deriving action plans. This position paper proposes a hybrid, statistical-formal approach toward achieving realtime autonomy.

Real-Time Nonlinear Attitude Control System for Nanosatellite Applications

This paper develops a fault-tolerant attitude controller for next-generation nanosatellites. The proposed fault-tolerant attitude control algorithms in this study are based on first-order and high-order sliding-mode control theories as well as fuzzy logic systems to achieve low cost and real-time autonomy. A locally asymptotically stable adaptive fuzzy first-order sliding-mode controller is chosen as the best solution to the local attitude control tracking problem. This novel fault-tolerant controller is validated by simulation results with reaction wheel Coulomb friction, saturation, noise, dead zones, bias faults, and external disturbances. Simulation and testing results presented in the paper demonstrate that the attitude control system can provide successful pointing and tracking in the presence of system uncertainties for a specified class of reaction wheel failures.