Sabiha Amin Wadoo

Feedback Control of Crowd Evacuation in One Dimension

This paper studies crowd models in one dimension. The focus of this paper is on the design of nonlinear feedback controllers for these models. Two different models are studied where dynamics are represented by a single partial differential equation (PDE) in one case and a system of hyperbolic PDEs in another, and control models are proposed for both. These include advective, diffusive, and advective-diffusive controls. The models representing evacuation dynamics are based on the laws of conservation of mass and momentum and are described by nonlinear hyperbolic PDEs. As such, the system is distributed in nature. We address the design of feedback control for these models in a distributed setting where the problem of control and stability is formulated directly in the framework of PDEs. The control goal is to design feedback controllers to control the movement of people during evacuation and avoid jams and shocks.

A LEGO based undergraduate control systems laboratory

The cost of establishing a traditional control systems laboratory usually runs into many thousands of dollars. This paper introduces an alternative method of teaching a control systems laboratory for undergraduate engineering students using LEGO NXT kits and ROBOTC software. The total cost of the kit and the software is under

Sliding Mode Control of Crowd Dynamics

350 which makes this combination a very cheap alternative for establishing a control systems laboratory. The set of experiments described here are ideal for colleges and universities that wish to introduce a control system laboratory curriculum at a minimal cost. In the first experiment the students observe and explore the working of the inbuilt ROBOTC PID controller for the LEGO NXT motor. In the second experiment, tuning of the PID controller assuming the system equations of the LEGO NXT motor system are unknown is studied. Ziegler-Nichols (Z-N) method and its Tyreus-Luyben (T-L) modification are also studied. In the third experiment, the transfer function of the LEGO NXT motor is derived using system identification by the experimental data modeling approach. PID control design using the new model is then finally studied in the fourth experiment using the experimentally obtained transfer function.

Optimal control of an autonomous underwater vehicle

In this paper, the design of nonlinear sliding mode controllers for models representing crowd dynamics in one dimension is presented. The main contribution of this paper is the stability analysis and robust control synthesis of hyperbolic partial differential equation (PDE) system models using the sliding mode method. The application of this research is in crowd control and in dynamically controlling the evacuation of pedestrians in the presence of disturbances. Crowd densities can change due to blocked exits or due to a varying influx of people. Recent advances in sensor technology have made the measurement of pedestrian densities and velocities possible. As such, the development and implementation of efficient control algorithms to control crowd movements that can avoid jams is realizable. The crowd model presented here is a system of nonlinear hyperbolic PDEs based on the laws of conservation of mass and momentum. The sliding mode control is designed in the presence of both matched and unmatched uncertainties due to external disturbance and parametric variations. The controllers designed are shown to be robust to disturbances.

Feedback ramp metering using Godunov method based hybrid model

This paper presents the optimal control of the kinematic model of an autonomous underwater vehicle. An optimal feedback control methodology is developed for the trajectory tracking of underwater vehicles. The motion of underwater vehicles is presented first as a kinematic model. For the design of feedback control, the system is linearized and transformed into a chained form. The control for trajectory tracking of the kinematic model is presented as an H-2 optimal design. Formulating LQG as H-2 optimization is useful as it can be generalized to include frequency-domain performance specifications.

Bayesian Safety Analyzer using multiple data sources of accidents

This paper presents a feedback control design for an isolated freeway ramp that utilizes a hybrid dynamical model for the traffic using Godunovs numerical technique. Feedback ramp metering designs in the past have relied on either discretized linearized method such as ALINEA, or nonlinear feedback designs based on ordinary differential equations for the traffic model. However, lumped parameter models fail to represent the rarefaction wave phenomenon of the distrbuted model. This paper uses Godunov based hybrid lumped model based on which feedback control design is proposed, and simulation results for the model are presented.

Sliding mode control of crowd dynamics with matched disturbance

In this study, the authors demonstrate the development of a Bayesian Safety Analyzer (BSA) using multiple accident data sources as well as traffic flow data. Simulations of the developed model are conducted using MATLAB FullBNT toolbox where different parameter estimations are used. This study demonstrates the efficiency as well as flexibility of using Bayesian analysis on large data sets containing a large number of attributes. The developed BSA can be used in the incident management process. It can assist decision makers in estimating the severity of a certain incident based on given attributes for better preparedness. BSA can also be used in order to assess the safety transportation system.

Establishing a cost effective embedded control and robotics engineering program optimal control of a two wheeled robot

In this paper the design of nonlinear sliding mode feedback controller for a model representing crowd dynamics is presented. The model is a system of partial differential equations based on the laws of conservation of mass and momentum. The equations of motion are described by a set of nonlinear hyperbolic partial differential equations. The feedback control is designed in presence of both matched and unmatched uncertainties due to external disturbance. The goal is to design a controller so as to minimize the effect of uncertainties on the movement of people. The control design method adopted is feedback linearization and sliding mode.

Feedback control of macroscopic crowd dynamic models

The cost of establishing a traditional control systems and robotics program usually runs into many thousands of dollars. As a result many undergraduate and graduate STEM institutions are unable to establish these important STEM programs in their curriculum. This paper introduces an alternative method of teaching important control system and robotics concepts using LEGO kits and ROBOTC and MATLAB software. This paper is the second in a series of papers that explore the extent to which LEGO kits can be used to introduce important control system and robotics concepts to both undergraduate and graduate students. Such a program has been successfully implemented in NYITs school of electrical and computer engineering. In this paper we discuss the control design of a two wheeled mobile robot. The dynamic model of the multi-input-multi-output (MIMO) system is nonlinear in nature. The nonlinear model is first linearized and then controlled by designing a full state feedback, LQR and H2 controllers. The system performance of each controller is simulated and compared with each other. Finally, the controllers are implemented and their performance is validated on the actual system. The establishment for embedded control system of two wheeled robot is design based on a MATLAB environment. Implementation of the software is completed using ROBOTC.

Time-Optimal Control for One Dimensional Evacuation System

This paper presents design of nonlinear feedback controllers for two different macroscopic models for two- dimensional pedestrian dynamics. The models presented here are based on the laws of conservation of mass and momentum. These models have been developed by extending one-dimension macroscopic vehicle traffic flow models that use two-coupled partial deferential equations (PDEs). These models modify the vehicle traffic models so that bi-directional controlled flow is possible. Both models satisfy the conservation principle and are classified as nonlinear, time-dependent, hyperbolic PDE systems. The equations of motion in both cases are described by nonlinear partial differential equations. We address the feedback control problem for both models in the framework of partial differential equations. The objective is to synthesize nonlinear distributed feedback controllers that guarantee stability of a closed loop system.