Abstract:
Pan Tilt Platform (PTP) is a complex electromechanical system whose dynamics is
characterized by coupled nonlinear behavior. The system is typically used as mounting platform for surveillance and tracking equipment. Mounted equipment as payload
requires accurate position and velocity control from PTP. Velocity control is required
in surveillance mode whereas position control is critical in tracking mode. Conventional controller design techniques rely primarily on making more robust controllers by
ignoring modeling details of physical system. In demanding applications, it’s hard to
design controllers for variable operational needs. The problem becomes difficult with
the presence of mechanical imperfections in the physical system. These mechanical
imperfections include mass imbalance in actuated bodies; friction and backlash at constraint joints of these actuated bodies. Practically, these mechanical imperfections are
always present in the physical system since they are difficult to remove.
NUST Man Portable Ground Surveillance Radar System is one such equipment
which has Pan Tilt Platform as its mounting and actuating platform. Current study is
aimed to model Pan Tilt Platform of the said system. The system requires position control in tilt axis; whereas position and velocity control in pan axis. The study aims at
building multiple PTP models having different abstraction levels and fidelity. It starts
with a multi-bond graph model of an ideal PTP using Object Oriented Modeling(OOM)
approach. OOM approach capture dynamics of individual components of multi-body
system and allows system modeling directly from system topology, thereby providing
modularity in the model. Ideal PTP model is made more realistic by adding details of
actuation mechanisms and mechanical imperfections in step wise manner. At first, a
payload as a mechanical imperfection of mass imbalance is introduced as an additional
rigid body in tilt axis. Then actuation details of DC servomechanism in pan joint and
linear actuator mechanism in tilt joint are added. Finally, mechanical imperfections of
ii
friction and backlash are added in pan joint. Friction is modeled using a general CSVS
friction model whereas backlash is modeled using classical dead zone model. Experiments are conducted on PTP of NUST Ground Surveillance Radar to quantify backlash
gap and to estimate friction parameters and Inertia. These parameters are estimated
using MATLAB Optimization Toolbox. For comparing individual models, their equivalent models are build using MSc ADAMS, Pro/Engineer and MATLAB/Simscape. In
the end, pan joint with classical friction and backlash models, is compared with experimental results.