MOTION SIMULATOR MODULE CYBERBALL
ABSTRACT
Introduction
to a new concept for motion simulation module
CYBERBALL. A simulator system with three
rotational degrees of freedom for unbounded bi-directional angular
displacement is described, in which the scenery as well as the
instrument panel and also the whole cockpit are simulated by digital
video graphics applying the novel VR technology by means of a VR head
mounted display and VR glove for virtual command. It follows a
description of selected topics of the applied system solutions, such
as axis drives, motion control, system kinematics and dynamics.
Keywords:
Flight Simulator, Virtual Reality, Automatic Control, Electrical
Drives, Interaction.
1 INTRODUCTION
A
motion simulation module Cyberball is an equipment capable of
submitting its user to mechanical sensations similar to the ones felt
when utilising the vehicle being simulated.
Flight
simulators are, no doubt, the most utilised ones and they allow the
physical reproduction of some sensations experimented during the
aircraft’s motion. Such simulators reached a high degree of
sophistication, reproducing accurately the cockpit’s
environment of a given aircraft aiming at training, specially for
taking off and landing operations, and to make the crew to become
acquainted to the on-board equipment in commercial aircrafts.
A
graphics processor generates the images of the flight scenery, which
are projected onto panels, in perfect synchronism with the mechanical
events of the interactive simulation process.
This
concept has been leading to the development of entirely specific
simulators for each model or version of aircraft, produced as
prototypes or in a small batch, requiring the development of
aeronautical and avionics equipment specially adapted for simulation
purposes, resulting in an extremely high cost equipment.
Consequently,
crews of small-sized aircraft usually do not have access to this kind
of training, which directly affects their performance and above all,
the flight safety.
Besides
its high cost, such simulators are also very limited at reproducing
the motions. This makes them inadequate for training acrobatic,
combat/fight and other manoeuvres, as well as for the class of
smaller-sized/higher dynamics aircrafts, such as helicopters.
Static
simulators, with very accessible costs, but limited to audio-visual
effects only, can even be used in introductory training phase, but
are not adequate in terms of hi-fi reproduction of flight sensations.
2 CONCEPTION OF A MOTION SIMULATOR SYSTEM
The
integration of new technological developments, specially in the field
of informatics and computers graphics and of electromechanical high
dynamic drives, helped the conception of an entirely new motion
simulator, whose patent has already been requested.
In this
conception, the rotating motions of the vehicle (aircraft, car, boat,
etc.. ) about the three mutually orthogonal Cartesian co-ordinate
axes are reproduced mechanically, as shown in FIG.1.

FIG.1
Conception of the simulator system.
The
mechanical conception of the simulator Cyberball consists of 4 basic
structural elements, starting with a stationary support base for the
module, and of 3 movable elements, one per co-ordinate axis, which
are a fork that rotates about A axis, an outer ring that rotates
about B axis articulated on the fork, and a cockpit (bubble) mounted
onto an inner ring, that rotates about C axis relative to the
external ring. Utilising these three axes of the simulator’s
plant, it can be reproduced the rotating motions about the three axes
that define the attitude of a vehicle: Y (yaw), P (pitch) and R
(roll). However, it is possible to show through kinematics analysis
that the dependency relation between the axes of the simulator’s
plant with the axes of the simulated vehicle do not remain constant.
Instead
of making physically available all the panel instruments, these are
also visually simulated through a graphics image processor in real
time, together with the operations scenery, and presented to the user
in the form of a special display mounted on a helmet-goggle for
virtual reality (FIG.2), which, whenever necessary, is
integrated to the piloting helmet utilised in the operation.

FIG.2 A
display helmet for virtual reality.
With
two small colour LCD screens of very high resolution, located in the
helmet, very near the user’s eyes, this display permits the
perfect illusion of 3-D stereoscopic images. Multiple orientation
(spacial angular position) sensors, installed in the helmet itself,
detect the motions of the user’s head, and inform the graphics
processor which portion of the scenery should be shown on the screen
at that instant, and in this way the user get a panoramic virtual
view of 360 degrees horizontally and vertically, which is identical
to the real view. Stereo sound is available through integrated
phones.
Only
the main devices for vehicle command are physically present, which
are, in the case of an aircraft, the control lever, the pedals, the
power command, and eventually the devices for special operations
command, e.g. for weapons. The buttons, keys and the other command
devices are also graphically simulated and actuated through
instrumented gloves for virtual reality, applied to the user’s
hands, whose positions in the 3-D space are continually informed to
the simulation processor, in such a way that a sufficient (virtual)
approximation to a certain command device will trigger its execution.
If the hands with gloves penetrate the instantaneous field of the
user vision, these will also be reproduced on the screens. There is
also inside the simulation cockpit a seat and safety devices for one
(or more) user(s), who are also exact replicas of the ones in the
vehicle being simulated.
With
these devices it is assured the full interaction between the user and
the system, which is a vital characteristic of virtual reality
systems, and distinguish them from other systems such as movies,
video, standard television, etc., that have a passive nature.
Thus, the
majority of the functions, which previously were carried out by
hardware, have been transferred to software solutions, that can be
replaced much more easily, quickly reconfiguring the simulator.
Therefore, in the same basic simulator module, with few alterations
to the command device (hardware) and by changing the simulation
program (the kinematics, dynamic models and their parameters), it is
effectively possible to simulate a variety of vehicles and
environments, either of similar nature or different.
It can
be noticed that this conception of a simulator, being preponderantly
generic, favours a series production with a very significant cost
reduction, which inclusively makes it viable its application as an
equipment for training users of land vehicles, both military and
civil, and also opening a perspective of utilisation as leisure
equipment, in a high level range, allowing to foresee a very
significant market for it.
3 THE SIMULATOR MODULE
Besides
the cost, also the weight of the movable parts of the simulator are
very much lower than the traditional equipment’s, which gives
it much higher dynamic characteristics, i.e. much greater realism in
the execution of the motions, and therefore a perfect illusion of
reality for the user, enabling also the simulation of radical
motions, such as acrobatics. The effects of the rectilinear
accelerations and displacements, which are difficult to simulate
mechanically in an exact way, are embedded in the graphic processing
of the scenery images, imposed to the user in synchronism with the
rotating motions.
The
rotating motions of the simulator module (FIG.1) result in
inertia variation on the three axes, which is a function of their
relative positions, superposed to varying gravity effects due to the
asymmetry of the movable part’s weight, and also related to
Coriolis and centrifugal effects.
The
specific dynamics of the vehicle being simulated needs to be imposed,
in real time and in a sufficiently precise way, to the own dynamics
of the simulator module.
In FIG.3 it is shown a block scheme of the main functions that
are part of the command and control system.

FIG.3
Scheme of the simulator command / control system.
The
calculation of the vehicle’s dynamic model is done from the
command variables, which are the signals originating from the command
elements available to the user (and eventually to an instructor) and
from the whole simulation program source (e.g. winds, amount of fuel,
road relief, interacting mobile objets, etc.). The result from the
dynamic model evaluation of the vehicle, which is applied to the
command of the simulator module, are the angular accelerations,
rotations and orientation of the vehicle along the three rotating
axes from their reference.
However,
since the simulated vehicle’s kinematics usually differs from
the simulator module kinematics, it is necessary to calculate an
inverse kinematics transformation, in order to transform the
variables (e.g. acceleration) of the vehicle into commands of each of
the movable parts of the plant, i.e. fork, external ring and cabine,
along the three axes of its reference.
In
order to synchronise the mechanical motion with the respective visual
graphical representations, the calculation of the dynamic model of
the vehicle requires information on the angular orientations where
the vehicle is forced by the imposed accelerations. Therefore, it is
necessary that, through another kinematics transformation, direct
transformation, the current orientations of the simulator module are
transformed back into those of the vehicle’s cabine.
The
synchronisation of the motions with the visual graphical part of the
simulation, requires that every information applied to the dynamic
model is also available to the graphics software, that manages the
co-ordinate system of the movable and geodesic elements.
4 DRIVE SYSTEM
It is
intended to generate tangible mechanical motions in the form of
accelerations and displacements capable of creating tactile
sensations similar to the ones that the user would experience if he
utilised the real vehicle. Consequently the reproduced motion must
correspond to the commands fed by the module’s command devices
and other variables generated in the simulation session program.
In order
to achieve that, the simulator module has a drive system with three
electromechanical servo-drives of high dynamics, one for each axis of
rotating movement. These are commanded co-ordinately by the motion
management processor (FIG.3), to which the calculated
acceleration references are imposed through the first kinematics
transformation.
The
motion manager performs many functions, as can be seen in FIG.3:
one of these is the calculation of the dynamic model of the
mechanical plant of the simulator module, in which, from the
acceleration, rotation and position references of the three axes
(generalised co-ordinates), the torque references for each of the
servo-drives are calculated.
It is
observed that the parameters of the dynamic model of the plant vary
with the orientation (attitude) and with the three orthogonal
components of rotation (speed) of the cabine: The necessary torques
for the acceleration are proportional to the inertia, which varies
with the attitude of the cabine; the required torques for overcoming
the forces due to the mutual influences of the motions of the three
movable parts (centrifugal and Coriolis effects) are functions of
both, attitude and rotations of the cabine; in turn, the torques
requested to overcome gravity are function of the cabine’s
attitude only.
Since
there is no direct relation between the three reference axes of the
vehicle’s cockpit, X, Y, Z (that correspond to those of the
simulator’s cabine), and the three driven physical axes of the
plant, qi, the execution e.g. of a pure yaw of
degrees requires co-ordinate ( simultaneous ) motions on the three
physical axes of the plant (FIG.4).

FIG.4
References of the simulator module. where: rotation about: Z
corresponds to yaw, Y to pitch; and X to roll of the cockpit and/or
vehicle; qi: rotations about the driven physical axes of the plant.
In
order to the cabine to follow the pilot’s, instructor’s
and/or scenery’s information determined motions, with the
specified rotational speeds and accelerations, the motion manager
must calculate the torque to be applied to each of the movable parts
of the plant by the respective drives. To calculate these torque, in
each situation, the manager receives as inputs the generalised
co-ordinates, i.e. the angles and their derivatives (rotational
speeds and accelerations) of each movable element, given relative to
the fixed reference base.
It is a
function of the inverse kinematics to execute the transformation of
the yaw, pitch and roll angles, which describe the vehicle’s
attitude, their rotations and accelerations, on the generalised
co-ordinates, i.e. on angles, rotations and accelerations of each of
the movable elements of the plant.
Once
the torque references are calculated by the simulator’s dynamic
model, these are then applied to three decentralised drive systems,
with independent control, one by rotating axis of the plant.
Other
functions of the motion manager correspond to full supervision of the
drives and the gathering of information on rotational speed based on
the position information of the sensors of the three axes.
5 IMPLEMENTATION
Every
transformations and calculations of the dynamic models are processed
in real time, with a calculation cycle that must be lower than 1ms.
The
dynamic model of the simulator is processed in real time by a high
performance DSP, hosted on a PC-like hardware with which it
communicates through a double-port RAM. ASICs for the interfacing of
the (angular) orientation measuring systems (encoders) of incremental
type, operate in parallel with the DSP. A DSP also calculates the
direct and inverse kinematics transformations.
The
calculation of the vehicle’s dynamic model is carried out by a
simulation manager processor that also controls the graphics
accelerator processor.
The
graphic processor (FIG.3) for real time image generation is
based on a graphic accelerator , and it can be utilised the
digitalisation of real scenery previously recorded, which allows for
instance the simulation of a flight procedure over a given ground,
during the day, night, under various meteorological conditions, etc.
Each
one of the three drive systems (FIG.5), is characterised as a
servo-system with an AC servo-motor (2) and a servo-converter with
vectorial control, commanded on torque, coupled with the special
reduction (3) with high reduction factor with negligible play and
coupled also to the angular position sensor (1) of the motor axis.
The digital control of each drive is done in closed loop, with a DSP
carrying out simultaneously the calculation of the control algorithms
of the three axes.

FIG.5
Simulator’s drive system.
For
energising as well as information transfer between the various
movable parts of the module, it is utilised slip ring and brush
devices. The information is transmitted via local network of high
transmission rate.
6 CONCLUSION
In
order to increase even more the realism of the simulations, the
command elements (control lever, etc.) may present tactile, acoustic,
etc. reactions corresponding to the real instantaneous dynamic state
of the motion. The cabine can be pressurised.
Simulators
of this kind can be interconnected via a network aiming at performing
simulation procedures with multiple participants interacting
simultaneously in real time, each of them with his/her own vehicle,
for e.g. race simulation, etc.
The new
conception also favours the future integration of new visualisation
technologies, drives, etc. and their application to still unexplored
activities.
A
down-sized instrumented model dynamic mock-up was built in order to
carry out advanced studies of the system’s kinematics (FIG.6).
The first prototype in natural size is being manufactured.
The
significant cost reduction achieved relative to state-of-the-art
simulators really will enable its application also as an equipment
for leisure, at least on the high end range, permitting to foresee a
very significant market for it.

FIG.6
Aspect of simulator’s small-sized driven mock-up.
|