We propose a new design for highly concurrent Internet services, whichwe call the staged event-driven architecture (SEDA). SEDA is intended

We have seen three important trends develop in the last five years in speech recognition. First, heterogeneous parameter sets that mix absolute spectral information with dynamic, or time-derivative, spectral information, have become common. Second, similariry transform techniques, often used to normalize and decor-relate parameters in some computationally inexpensive way, have become popular. Third, the signal parameter estimation problem has merged with the speech recognition process so that more sophisticated statistical models of the signal’s spectrum can be estimated in a closed-loop manner. In this paper, we review the signal processing components of these algorithms. These al-gorithms are presented as part of a unified view of the signal parameterization problem in which there are three major tasks: measurement, transformation, and statistical modeling. This paper is by no means a comprehensive survey of all possible techniques of signal modeling in speech recognition. There are far too many algorithms in use today to make an exhaustive survey feasible (and cohesive). Instead, this paper is meant to serve as a tutorial on signal processing in state-of-the-art speech recognition systems and to review those techniques most commonly used. In keeping with this goal, a complete mathematical description of each algorithm has been included in the paper.

As Internet services become more popular and pervasive, a critical problem that arises is managing the performance of services under extreme overload. This paper presents a set of techniques for managing overload in complex, dynamic Internet services. These techniques are based on an adaptive admission control mechanism that attempts to bound the 90th-percentile response time of requests flowing through the service. This is accomplished by internally monitoring the performance of the service, which is decomposed into a set of event-driven stages connected with request queues. By controlling the rate at which each stage admits requests, the service can perform focused overload management, for example, by filtering only those requests that lead to resource bottlenecks. We present two extensions of this basic controller that provide class-based service differentiation as well as application-specific service degradation. We evaluate these mechanisms using a complex Webbased e-mail service that is subjected to a realistic user load, as well as a simpler Web server benchmark.

A widely used approach to achieving service level objectives for a target system (e.g., an email server) is to add a controller that manipulates the target system's tuning parameters. We describe a methodology for designing such controllers for software systems that builds on classical control theory. The classical approach proceeds in two steps: system identification and controller design. In system identification, we construct mathematical models of the target system. Traditionally, this has been based on a first-principles approach, using detailed knowledge of the target system. Such models can be di#cult to build, and too complex to validate, use, and maintain. In our methodology, a statistical (ARMA) model is fit to historical measurements of the target being controlled. These models are easier to obtain and use and allow us to apply control-theoretic design techniques to a larger class of systems. When applied to a Lotus Notes groupware server, we obtain model fits with R no lower than 75% and as high as 98%.

A method is presented for generating shaped command inputs which significantly reduce or eliminate endpoint vibration. Desired system inputs are altered so that the system completes the requested move without residual vibration.

This paper describes research done at the Artificial Intelligence Laboratory of the Massachusetts Institute of Technology. Support for the Laboratory's artificial intelligence research is provided in part by the Advanced Research Projects Agency of the Department of Defense under Office of Naval Research contract N00014-89- J-3202, and in part by the National Science Foundation grant MIP-9001651. The author is also supported by a G.Y. Chu Fellowship

Computational models have emerged as a powerful tool for studying the complex task of driving, allowing researchers to simulate driver behavior and explore the parameters and constraints of this behavior. In this paper we investigate the advantages of developing rigorous computational models of driver behavior in cognitive architectures — computational frameworks with underlying psychological theories that incorporate basic properties and limitations of the human system. In particular, we describe an integrated driver model developed in the ACT-R cognitive architecture and demonstrate how this model accounts for the steering profiles, lateral-position profiles, and gaze distributions of human drivers during lane keeping, curve negotiation, and lane changing. The model has implications both for theoretical accounts of complex dynamic tasks in the context of cognitive architectures and for practical applications in predicting and recognizing driver behavior and distraction.

Many real-world systems are naturally modeled as hybrid stochastic processes, i.e., stochastic processes that contain both discrete and continuous variables. Examples include speech recognition, target tracking, and monitoring of physical systems. The task is usually to perform probabilistic inference, i.e., infer the hidden state of the system given some noisy observations. For example, we can ask what is the probability that a certain word was pronounced given the readings of our microphone, what is the probability that a submarine is trying to surface given our sonar data, and what is the probability of a valve being open given our pressure and flow readings. Bayesian networks are

Abstract — An inevitable collision state for a robotic system can be defined as a state for which, no matter what the future trajectory followed by the system is, a collision with an obstacle eventually occurs. An inevitable collision state takes into account both the dynamics of the system and the obstacles, fixed or moving. The main contribution of this paper is to lay down and explore this novel concept (and the companion concept of inevitable collision obstacle). Formal definitions of the inevitable collision states and obstacles are given. Properties fundamental for their characterisation are established. This concept is very general and can be useful both for navigation and motion planning purposes (for its own safety, a robotic system should never find itself in an inevitable collision state!). The interest of this concept is illustrated by a safe motion planning example.

Distributed real-time and embedded (DRE) applications have become critical in domains such as avionics (e.g., flight mission computers), telecommunications (e.g., wireless phone services), tele-medicine (e.g., robotic surgery), and defense applications (e.g., total ship computing environments). DRE applications are increasingly composed of multiple systems that are interconnected via wireless and wireline networks to form systems of systems. A challenging requirement for DRE applications involves supporting a diverse set of quality of service (QoS) properties, such as predictable latency/jitter, throughput guarantees, scalability, 24x7 availability, dependability, and security that must be satisfied simultaneously in real-time. Although a growing number of DRE applications are based on QoS-enabled commercial-o#-theshelf (COTS) hardware and software components, the complexity of managing long lifecycles (often years) remains a key challenge for DRE application developers. For example, substantial time and e#ort is spent retrofitting DRE applications when their COTS technology infrastructure changes.