ENGG4420: LECTURE 3 ‐‐ CHARACTERISTICS OF REAL‐TIME SYSTEMS TIME CONSTRAINTS: every real‐time task is associated with some time constraints: a) Deadline associated with tasks ‐‐ common constraint; b) Delay; c) Duration A. It is the responsibility of the Real‐Time Operating System (RTOS) to ensure that all tasks meet their respective time constraints. A logically correct result produced after the deadline would be considered incorrect. a. NEW CORRECTNESS CRITERION: correctness in real‐time system implies both logical correctness of results, and correctness of time at which the results are produced. B. Safe system is one that does not cause any damage even when it fails a. Reliable system is one that can operate for long durations of time without any failure b. SAFETY‐CRITICALITY: In many real‐time systems the issues related to safety and reliability are bounded together to generate safety‐critical systems ‐‐ safety‐critical system is required to be highly reliable. C. CHAPTER 1 BY RADU MURESAN Page 1
Transcript
ENGG4420: LECTURE 3 ‐‐ CHARACTERISTICS OF REAL‐TIME SYSTEMSTIME CONSTRAINTS: every real‐time task is associated with some time constraints: a) Deadline associated with tasks ‐‐ common constraint; b) Delay; c) Duration
A.
It is the responsibility of the Real‐Time Operating System (RTOS) to ensure that all tasks meet their respective time constraints.
A logically correct result produced after the deadline would be considered incorrect.
a.
NEW CORRECTNESS CRITERION: correctness in real‐time system implies both logical correctness of results, and correctness of time at which the results are produced.
B.
Safe system is one that does not cause any damage even when it fails
a.
Reliable system is one that can operate for long durations of time without any failure
b.
SAFETY‐CRITICALITY: In many real‐time systems the issues related to safety and reliability are bounded together to generate safety‐critical systems ‐‐ safety‐critical system is required to be highly reliable.
C.
CHAPTER 1 BY RADU MURESAN Page 1
Sensor data may be sensed periodically, aperiodically ‐‐ therefore real‐time systems must process data from all the sensors concurrently, otherwise signals may be lost and the system may malfunction
a.
These systems can be considered non‐deterministic, since the behaviour of the system depends on the exact timing of its inputs
b.
A non‐deterministic computation is one in which two runs using the same set of input data can produce two distinct sets of output data.
c.
CONCURRENCY: a real‐time system usually needs to respond to several independent events within very short and strict time bounds
D.
STABILITY ‐‐ under overload conditions, real‐time systems need to continue to meet the deadlines of the most critical tasks, though the deadlines of non‐critical tasks may not be met
E.
CHAPTER 1 BY RADU MURESAN Page 2
The different events of interest arise at geographically separated locations
○
There is need to handle locally the events and responses produced to them to prevent overloading of the communication network
○
Therefore, the sensors and the actuators may be located at places where the events are generated ‐‐ example is a petroleum refinery plant.
○
Many distributed as well as centralized real‐time systems have a feedback structure as shown below
○
DISTRIBUTED AND FEEBACK STRUCTURE ‐‐ in many real‐time systems, the different components of the system are naturally distributed across widely geographic area.
F.
CHAPTER 1 BY RADU MURESAN Page 3
Criticality of a task is determined by examining how critical are the results produced by the task to the proper functionality of the system.
○
The higher the criticality of a task the more reliable it should be made
Criticality of tasks needs to be taken in consideration when we design for fault‐tolerance
○
If a highly critical task fails, immediate failure detection and recovery are important
○
Note that task priority is a different concept then task criticality and the priority of a task should not be determined solely based on its criticality
○
TASK CRITICALITY: is a measure of the cost of failureof a task
G.
Ex: in a chemical plant that is setup to work non‐stop taking corrective actions on a failure becomes difficult in the absents of operators
○
Even if corrective actions cannot be immediately taken it is important that a failure does not result in a catastrophic situation.
○
A failure needs to be detected and the system should continue to function in a graceful degraded mode, rather then shut down.
○
EXCEPTION HANDLING ‐‐many real‐time systems work round‐the‐clock.
H.
CHAPTER 1 BY RADU MURESAN Page 4
Traditional systems compute the output data as some function of the input data
○
In each interaction step, the results computed are used to carry out some actions on the environment.
The reaction of the environment is sampled and is fed back to the system ‐‐ the computations in real‐time system can be considered non‐terminating.
In contrast real‐time systems do not produce any output data but enter into an ongoing interaction with their environment
○
REACTIVE: a reactive system is one in which an ongoing interaction between the computer and the environment is maintained.
Ex: cell phones use various custom components, the MPFI use specialized embedded processor.
○
CUSTOM HARDWARE: a real‐time system is often implemented on custom hardware that is specifically designed and developed for the purpose.
CHAPTER 1 BY RADU MURESAN Page 5
Connection to the environment through a wide range of interface devices, and
Receive and send variety of stimuliExample of an embedded system is the Multi‐Point Fuel Injection (MPFI) system.
Common features of real‐time systems and embedded computers are:
○
REAL‐TIME EMBEDDED COMPUTERS ‐‐ are computers that implement one or more functional elements of a real‐time system and is not a stand‐alone computing machine
Complex due to its environment interaction○Difficult to test during the design process○Need to use plant models and environment models in order to develop a functional design
○
LabView is a program that is very useful in simulating environment, modeling plants, testing, and implementing support software.
○
Majority of the 41x designs that I supervise have made use of the LabView in order to show proof of concept for their embedded applications (real‐time or not).
○
DESIGN OF REAL‐TIME EMBEDDED SYSTEMS
CHAPTER 1 BY RADU MURESAN Page 6
CLASSIFICATIONS OF REAL‐TIME SYSTEMS BASED ON SYNCHRONIZATION
Tasks are internal actions carried by the computer○External processes have their own time scale○
Tasks are connected by physical devices to external processes. We need to consider that:
•
Real‐time operation is an action carried out in the computer and relates to the time scales of the external processes
•
Operations are carried out according to a time scheduleClock based ‐‐ so we have clock based systems○
Event based ‐‐ the events determine the synchronization
○
Interactive ‐‐ a set of operations should be completed within a predetermined time.
○
Synchronization types between external processes and the tasks (internal actions) are:
•
IMPORTANCE OF PLANT CHARACTERISTIC
Time constants determine the sampling rate○
The time constant of a plant is a measure of the time taken by the plant to respond to a change in input or load and is called the plant characteristic
•
CHAPTER 1 BY RADU MURESAN Page 7
Real‐time requirements ‐‐ computer used for control must be able to carry out all the required operations (i.e., measurement, control, actuation) within each sampling interval.
•
Synchronization is obtained by adding a real‐time clock to the computer system and using interrupts from this clock.
•
The task is to run once per period T, or it is to run at exactly T unit intervals
○
CLOCK BASED TASKS ‐‐ are referred to as cyclic or periodic
Ex: turning off a pump, closing a valve, or switching a motor off in response to the closure of a position microswitch).
Event based systems are also used extensively to indicate alarm conditions and initiate alarm actions in industrial processes
Actions need to be performed in response to some event
○
The general real‐time requirement for event‐based systems is that the system responds within a given maximum time to an event.
○
EVENT BASED TASKS ‐‐ also called aperiodic
CHAPTER 1 BY RADU MURESAN Page 8
Interrupts ‐‐ to inform the computer that action is required
○
Polling ‐‐ the computer periodically asks (polls) the various sensors to see if action is required.
○
Implementation of event based systems•
Example: a task may be required to start within 0.5 seconds or produce an output within 0.5 seconds of an event occurring
Such tasks may have deadlines expressed in terms of start or finish times (or even both).
○
Events based tasks are called aperiodic tasks ‐‐events occur at non‐deterministic intervals
•
The real‐time requirement for these system is usually expressed in terms such as: the average response time should not exceed 20 seconds
○
INTERACTIVE SYSTEMS ‐‐ reservation system for airlines, hotels, etc. Or a bank teller system.
Interactive versus event based ‐‐ interactive systems respond at a time determined by the internal state of the computer and without reference to the environment.
•
Interactive versus clock based ‐‐ clock based are tightly synchronized to an external process through a clock while interactive are not.
•
CHAPTER 1 BY RADU MURESAN Page 9
TYPES OF REAL‐TIME SYSTEMS AND TASKSUsing the deadline constraint we can have two major categories of real‐time tasks: hard real‐time and softreal‐time.
•
Hard, soft, or firm real‐time task○
Some authors divide the real‐time tasks based on the consequences of a task missing a deadline in:
•
Within a real‐time system it is not necessary to have all task belonging to the same category
•
The system is considered to have failed whenever any of its hard real‐time tasks does not produce its required results before their deadlines.
○
Robot system ‐‐ a robot cyclically carries out various tasks such as communication with the host, logging all completed activities, sensing the environment to detect any obstacle, tracking the object of interest, path planning, effecting next move.
Consider that a robot suddenly encounters an obstacle ‐‐ the robot needs to detect it and try to escape colliding with it quickly ‐‐ if concerned tasks are not completed in time then collision takes place and the robot would be considered to have failed.
Examples:○
HARD REAL‐TIME TASKS ‐‐ it is a task that is constrained to produce its results within certain predefined time bounds
CHAPTER 1 BY RADU MURESAN Page 10
OTHER examples of hard real‐time systems is the anti‐missile system, etc.
•
HOMEWORK: 1) describe why the anti‐missile system is a hard real‐time system; 2) think of other examples that are real‐time hard system and why?In practical system the time bounds for hard real‐time tasks usually range from several micro‐seconds to a few milliseconds.
•
There is no reward in completing it ahead of its deadline
○
It should be noted that a hard real‐time task does not need to be completed within the shortest possible time but within the specified time bound
•
FIRM REAL‐TIME TASKS ‐‐ the results of a firm real‐time task is associated with a predefined deadline, but unlike a hard real‐time task if a firm task does not complete before its deadline the system does not fail (the late results are discarded)
CHAPTER 1 BY RADU MURESAN Page 11
Firm real‐time tasks are typically encountered in multimedia applications
•
The system takes pictures of an enemy territory and beams it to a ground station computer frame by frame
The ground computer proceeds to find positional difference of various objects of interest with respect to previous frames
When the computer is overloaded a new image could be received before an old one has been processed ‐‐ in this case the old one can be discarded
Satellite‐Based Tracking of Enemy Movements○
Video Conferencing ‐‐ HOMEWORK ‐‐ explain why this application is a firm real‐time system.
○
Examples:•
The associated time bounds for firm real‐time tasks range typically from a few milliseconds to several hundreds of milliseconds.
•
SOFT REAL‐TIME TASKS: these tasks also have time bounds associated with them but the timing constraints of soft real‐time tasks are not expressed as absolute value ‐‐ instead, the constraints are expressed in terms of the average response times required.
CHAPTER 1 BY RADU MURESAN Page 12
Once a request for reservation is made, the response should occur within 20 seconds on an average
Response result may be in the form of a printed ticket or an apology on an unavailability ticket
Alternatively, we might state the constraint on the ticketing task as: at least in case of 95% of reservation requests, the ticket should be processed and printed in less than 20 sec.
Railway seat reservation system○
Using the above example ‐‐ if the ticket is printed in about 20 sec, we feel that the system is working fine and have a feel of obtaining instant results
Missed deadlines of soft real‐time tasks do not result in system failure. However, the utility of the result produced by a soft real‐time task falls continually with time after the deadline expired
IMPACT of the deadline failure in a soft real time task○
Examples of Soft Real‐Time Systems•
CHAPTER 1 BY RADU MURESAN Page 13
In practical applications, the time bounds for typical soft real‐time tasks usually range form a fraction of a second to a few seconds
•
NON‐REAL‐TIME TASKS: a non real‐time task is not associated with any time boundsMost interactive computations we perform nowadays are handled by soft‐real time tasks.
•
A few examples of non‐real time tasks could be batch processing jobs, e‐mail, background tasks, etc.
•
But, we could argue that even these tasks could be considered soft real‐time to some degree
•
WHAT IS THEN THE DIFFERENCE BETWEEN a non‐real time tasks and a soft real‐time tasks??For non‐real‐time tasks the associated time bounds are typically of the order of a few minutes, hours, or even days.
•
In contrast for the soft real‐time tasks the time bounds are the most of the order of a few seconds.
•
CHAPTER 1 BY RADU MURESAN Page 14
ENGG4420. CHAPTER 1: Real-Time Computer Control. Developed by Radu Muresan, Univesity of Guelph 17
Hard Soft
Periodic(Cyclic)
A periodic(Event)
Periodic(Cyclic)
A periodic(Event)
ee Tit )(atit sc )(
n
i
asc tit1
)(n
1
stTn /
n
i
ae Tit1
)(n
1
stTn /
tc(i) the interval between the i and i – 1 cycles,te(i) the response time to the ith occurrence of event e,ts the desired periodic (cyclic) interval,Te the maximum permitted response time to event e,Ta the average permitted response time to event e measured over
some time interval T,n the number of occurrences of event e within the time interval T,
or the number of cyclic repetitions during the time interval T,a a small timing tolerance.
FORMAL CLASSIFICATION OF REAL‐TIME TASKS
CHAPTER 1 BY RADU MURESAN Page 15
CLASSIFICATION OF PROGRAMS
Real‐time tasks and non‐real‐time tasks.○Subdivide the real‐time tasks into hard and soft○
It is important to separate the computer control activities into:
•
The division of software into small, coherent modules is an important design technique.
•
Verification of the applications ...
Sequential ‐‐ actions are strictly ordered as a time sequence. The behaviour of the program depends only on the effects on the individual actions and their order.
○
Processes or tasks are partly sequential but are executed concurrently ‐‐ communicate through shared variables and synchronization signals.
Verification of the applications ...
Multi‐tasking ‐‐ the actions required to perform are not necessarily disjoint in time (concurrency). Such programs can be built from a number of parts called processes or tasks.
○
Verification ‐‐ need models of environment.
Real‐Time ‐‐ in addition to its action not necessarily being disjoint in time, the sequence of some of its action is determined by the environment
DESIGN EXAMPLE ‐‐ REAL‐TIME COMPUTER CONTROL FOR HOT‐AIR BLOWER SYSTEM
The approach to the design of real‐time control systems is no different in outline from that required for any computer‐based system.
○
The planning phase ‐‐ is concerned with interpreting user requirements to produce a detailed specification of the system to be developed and an outline plan of resources, people, time, equipment, costs‐required to carry out the development. At this stage preliminary decisions regarding the division of functions between hardware and software will be made ‐‐ The outcome of this stage will be a specification or requirement document.
The development phase ‐‐ is composed of various stages such as: preliminary stage, design stage, and testing stage. During these stages there is need for extensive cooperation between hardware and software designers.
The work can be divided into two main sections:○
CHAPTER 1 BY RADU MURESAN Page 18
HOT AIR BLOWER PLANT ‐‐MAIN PLANT OF THE SYSTEM EXAMPLE
BridgeCircuit
& Amplifier
FullyOpen
ThyristorUnit
TubeAir Flow
PowerInput
A
Computer
B
TemperatureMeasurement
(0-10V)
Auto Man
Air Inlet Heater
ManualFullyClosed
Air Inlet Position
On/OffClockwise/
Anticlockwise
Blower
MotorControl
Operator
VariableAir Inlet
centrifugal fan (blower),○thermistor,○heating element,○reversible motor,○potentiometer wiper for inlet air,○microswitches,○slider potentiometer for the reference temperature,○operator panel.○
This plant example (a hot‐air blower) is presented in order to illustrate the various operations of a computer control system. As seen from the diagram, the components of the hot‐air blower plant are:
CHAPTER 1 BY RADU MURESAN Page 19
COMPONENTS DESCRIPTIONCentrifugal fan ‐‐ blows air over a heating element and into a tube.
•
The thermistor bead ‐‐ placed at the outlet end of the tube forms one arm of a bridge circuit (a common configuration used to measure small resistor variations). The amplifier output of the bridge circuit is available at B and provides a voltage, in the range 0 to 10 volts, proportional to temperature.
•
The current supplied to the heating element can be varied by supplying a DC voltage in the range 0 to 10 volts to point A.
•
The position of the air‐inlet cover to the fan is adjusted by means of a reversible motor (can be adjusted by the operator or computer). The motor operates at constant speed and is turned on or off by a logic signal applied to its controller; a second logic signal determines the direction of rotation.
•
A potentiometer wiper is attached to the air‐inlet cover and the output voltage is proportional to the position of the cover.
•
Microswitches are used to detect when the cover is fully open and fully closed.
•
in manual mode the heat output and fan cover position can be adjusted using potentiometers. Switches are provided to operate the fan and heater
•
Panel lights indicate fan on, heater on, cover fully open, cover fully closed, and auto/manual status
•
The operator panel: automatic and manual control•
The desired output temperature is known as the set point and is set by the operator using a slider potentiometer. The computer can read this setting.
•
The question is: how do we implement a system like this?•
CHAPTER 1 BY RADU MURESAN Page 20
FEATURES OF THE PLANTPanel lights: fan on; heater on; cover fully open; cover fully closed; and auto/manual status
•
Set point for the control system ‐‐ the desired output temperature is set by the operator and it can be read by the computer.
•
In manual mode the heat output and fan cover position can be adjusted by the operator using potentiometers.
•
Switches are provided to operate the fan and the heater.
•
COMPUTER CONTROL FEATURES
Monitoring of the plant○Control of the plant○Actuation of the plant○And communication to the operator○
The operation of the plant requires that software be provided to support:
•
CHAPTER 1 BY RADU MURESAN Page 21
COMPUTER CONTROL OF A HOT‐AIR BLOWER
Digital Input
ADC
Computer
DAC Digital Output
BridgeCircuit
Air InletPosition
Heater Circuit
Motor Control
OperatorPanel
AirFlow
Thermistor
AirInlet
Close
AirInlet Open
Blower
HeaterElement
DirectionAuto/Manual
ToOperator
Panel
On/Off
AirInlet
Air Inlet Motor
A general schematic of the system is shown above. The computer interacts with the digital input and output interfaces and with the ADC and DAC. The interfaces are directly connected to the plant. Complex micro‐processors incorporate many I/O interfaces on the same chip with the ALU. Many of these interfaces are part of the embedded processor.
CHAPTER 1 BY RADU MURESAN Page 22
MONITORING FUNCTIONMonitoring involves obtaining information about the current state of the plant.
•
Analog signals: air temperature, fan‐inlet cover position, temperature reference.
○
Digital (logic) signals: fan‐inlet cover position (i.e., fully open, fully closed); status signals (i.e., auto/manual, fan motor on, heater on).
○
In our example, the information is available from the plant instruments in the following two forms:
•
ACTUATION REQUIREMENTSThe provision of a voltage proportional to the demanded heat output to drive the heater control;
•
Logic signals indicating on/off and the direction in which the fan‐inlet cover is to be moved;
•
Logic signals for the operator display.•
CHAPTER 1 BY RADU MURESAN Page 23
CONTROL ‐‐ involves parallel logic operations, time‐sequential control and timing of operation. Specifically we need to implement:
The temperature, and ○The position of the fan‐inlet cover.○
The digital equivalent of continuous feedback control (direct digital control DDC) for control of:
•
For example, the heater should not be on if the fan is not running.
○
Sequence and interlock control operations are also required
•
The change over needs to be done without disturbing the temperature of the air at the output of the tube ‐‐ bumpless transfer.
○
The computer needs also to handle automatic change‐over from simply tracking (monitoring) the manual control operations to controlling the system when the operator requests a change from manual to automatic control
•
BUMPLESS TRANSFER ‐‐ a change from manual to auto mode which does not cause disturbance in the plant output.
