Business analytics | EPFL Graph Search

Business analytics (BA) refers to the skills, technologies, and practices for iterative exploration and investigation of past business performance to gain insight and drive business planning. Business analytics focuses on developing new insights and understanding of business performance based on data and statistical methods. In contrast, business intelligence traditionally focuses on using a consistent set of metrics to both measure past performance and guide business planning. In other words, business intelligence focusses on description, while business analytics focusses on prediction and prescription. Business analytics makes extensive use of analytical modeling and numerical analysis, including explanatory and predictive modeling, and fact-based management to drive decision making. It is therefore closely related to management science. Analytics may be used as input for human decisions or may drive fully automated decisions. Business intelligence is querying, reporting, online ana

Stabilization of electrodynamic bearings with active magnetic dampers

Qingwen Cui

Electrodynamic Bearings (EDBs) are a kind of passive magnetic bearings that exploits the interaction between the induced eddy currents in a conductor and a magnetic field to provide re-storing forces. They have been regarded as an appealing alternative to Active Magnetic Bearings (AMBs), having the ability to provide positive stiffness passively without introducing negative stiff-ness in any direction. Compared to AMBs, EDBs present advantages such as lower cost and higher reliability due to simplicity of configurations. One of the most interesting features of EDBs is the possibility to obtain stable levitation using standard conductive materials at room temperature, requiring no control systems, power electronics or sensors. Thus EDBs could be suitable solutions for highâspeed rotating machinery such as flywheels, small size compressors, centrifuges and vac-cum pumps. Despite these promising characteristics of EDBs, applications are still limited because of instability issues. The main problem is that the effect of the rotating damping force in EDBs causes unstable behavior of the rotor. In existing solutions, stabilization is achieved mainly by introducing non-rotating damping to the rotor with passive ways. Although stable levitation is possible, the effectiveness of the existing methods is still limited. A hybrid solution has been proposed in this thesis, where EDBs are combined with active magnetic dampers (AMDs). Using similar magnetic actuators as those used in classical active magnetic bearings (AMBs), nonârotating damping forces are applied on the rotor supported by EDBs to obtain stable operation. This system is designed to exploit the high reliability of EDBs, overcoming the stability problem by means of controllable AMDs. It results in increased global system reliability. In case of AMBs failure, the EDBs are able to guarantee a stable levitation down to a certain speed considered safe for touchâdown. During the operation speed range, the AMDs provide nonârotating damping to stabilize the rotor. This nonârotating damping can be easily tuned during rotor operation phase. At low speeds when the EDB forces are not sufficient to support the rotor, the active magnetic actuators work as AMBs to guarantee stable levitation of the rotor in a wide speed range. Besides, the EDBâAMD configuration also allows characterizeing in dynamic condition, which opens the possibility to establish damping strategy that can in perspective be implemented by totally passive means, such as eddy currents, elastomeric mounts. The combination of EDB and AMD forces are studied both analytically and experimentally. An analytical model of the system, as well as a test rig, has been built. Simulations and experi-mental tests validate the model and characterize the system. The effectiveness of the proposed solution is confirmed. The control strategy of AMDs and stabilizing alternatives of EDBs are dis-cussed consequently. The combination of EDB and AMD can be exploited to investigate easily dif-ferent damping strategies.

EPFL2016

An Analytical Model for the Contention Access Period of the Slotted IEEE 802.15.4 with Service Differentiation

Giovanni De Micheli, Nadia Khaled, Eugène David Ngangue Ndih

The IEEE 802.15.4 standard is poised to become the global standard for low data rate, low energy consumption wireless sensor networks (WSN). By assigning the same sets of contention access parameters for all data frames and nodes, the contention access period (CAP) of the slotted IEEE 802.15.4 medium access control (MAC) currently provides a priorityindependent channel access functionality and no service differentiation. Several recent WSN applications such as wireless body sensor networks, however, may require service differentiation and traffic prioritization support to accommodate potential high priority traffic (e.g., alarms or emergency alerts). By allowing different sets of access parameters and data frame lengths for different priority classes, this paper develops a Markov-chain-based analytical model of the CAP of the IEEE 802.15.4 MAC with service differentiation, under unsaturated traffic conditions. In particular, given two priority classes, our analytical model is used to evaluate the performance of a simple, yet effective, contention-window-based service differentiation strategy, in terms of the resulting throughput, average frame service time and access priority for each priority class. The accuracy of the analytical model is validated by extensive ns-2 simulation.

2009

Haptic instrumentation for an interventional radiology simulator

Dejan Ilic

Low trauma, reduced costs and fast recovery are only a few factors why minimally invasive surgery (MIS) is taking over classical surgical methods. Interventional Radiology (IR) is a minimally invasive procedure where thin tubular instruments are steered through the patient's vascular system under X-ray imaging. In order to perform these procedures, a radiologist has to be trained to master hand-eye coordination, instrument manipulation and procedure protocols. The importance of training IR procedures rises with the increasing variety and sophistication of the procedures performed using this approach. Apart from the traditional "see one, do one, teach one" model, radiologists have only had few mock devices with limitations in practical use. Training on animals is an alternative, but its use is hindered by high costs and ethical issues. The existing simulation systems all have major drawbacks: the requirement to use modified instruments, high inertia of the haptic feedback that creates a noticeably degraded dynamic behavior, or excessive friction that is not properly compensated. In this thesis we propose a quality training environment dedicated to IR. The system is composed of a virtual reality (VR) simulation of the patients anatomy linked to a robotic interface providing haptic force feedback. This thesis focuses on the requirements, design and prototyping of a specific haptic interface for the manipulation of catheters and guide wires. Translational tracking and force feedback on the catheter are provided by two cylinders forming a friction drive arrangement. The complete friction drive can be set in rotation with an additional motor providing torque feedback. A force and a torque sensor are integrated in the cylinders for direct measurement on the catheter enabling disturbance cancellation with a closed-loop force control strategy. One of the specific requirements during a simulated intervention is to measure the position of the tip of a guide wire inside a catheter. This is a so called "tube in tube" problem. The guide wire is detected at the entrance of the simulator and then grasped by a micro-gripper. Forces and movements of the guidewire moving inside the catheter are then transmitted through a gripper and along a rod to an external unit. This thesis also reports a study of several micro-gripper designs developed and tested on a haptic force feedback interface developed at the LSRO-EPFL. A virtual reality environment has been developed to provide real-time visualisation and force feedback. Several different software routines are integrated in the system. During an IR procedure the tip of an employed instrument interacts with a wall of blood vessels and a repulsive force is transmitted to the user. Force feedback helps radiologist to orient and navigate the inserted instrument in the vessel. This thesis presents a novel method for measurement of internal constraints during an IR procedure. The measurements help to determine the maximum interaction forces between the catheter and the wall of the blood vessel and to observe the shape of dynamic interaction during manipulation. In order to recreate a realistic simulation environment it is necessary to derive the elastic properties and behavior of arteries when contact with an instrument occurs. A novel method for measurement of mechanical properties of blood vessels is proposed in this work. Information derived with this method is also very useful to determine coherent boundary conditions for hemodynamic simulations. Measurements of the modulus of elasticity of an artery are carried out by measuring the deformations due to the inflation of an angioplasty balloon catheter used for IR procedures. A test bench, consisting of an angioplasty balloon, a PVA model and an actuator used to inflate the balloon, is developed for realization of experiments. The pressure-volume curve during the inflation of the balloon is observed. The Young's modulus is derived with an analytical model of the measurement system. The results are then analyzed and compared to existing data from literature.

EPFL2005