‫بسم اهلل الرحمن الرحیم‬

Embedded and Real-Time Systems

Spring 2025

Elham Cheshmikhani
[email protected]

Department of Computer Science & Engineering

Shahid Beheshti University
Copyright Notice

This lecture is adopted from

- Real-Time Embedded Systems Course 2019, Alireza Ejlali,
Sharif University of Technology
- Real-Time Embedded Systems Course 2020, Hossein Attarzadeh,
Beheshti University

Marwedel’s and LeeSeshia’s books, Chapter 2

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 2
Recap

• What is an embedded system?

• Differences between embedded systems and cyber physical

systems

• Real-time embedded systems

• Requirements
– Real-time-ness, efficiency, dependability, and extra-function requirements

• Challenges of implementation in pure hardware and software

– Modeling is the best choice

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 3

Content

Let’s model an embedded system

• Capturing the requirements for modeling techniques

• Providing an overview of models of computation (MoC)

– Von Neumann languages

– Modeling continuous dynamics – differential equations

• Feedback control systems – time-domain modeling

– Modeling discrete dynamics – finite-state machines (FSM)

– Modeling hybrid systems – timed automata

– Concurrent models of computation

• Synchronous composition

• Dataflow models

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 4

Embedded Systems

• Information processing systems embedded into a

larger product
• Main reason for buying is not information processing

• Ubiquitous Computing:
– Information anytime, anywhere
• Embedded systems provide fundamental technology

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 5

Simplified Design Fellow

• Embedded system design chart

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 6

Reactive Systems

• Reactive Systems = Event-based Systems

• The traditional paradigms of programming (i.e.
model of computable functions) are inappropriate
– Model of computable functions
• Von Neumann paradigm
• Sequential computing
• Suitable model for reactive systems:
– Automata-based programming paradigm

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 7

Model

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 8

• Embedded system design chart

• The first step for design is specification for the model

of the system under design (SUD)
• The model contains exactly those characteristics and
properties of the modeled entity that are relevant for a
given task

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 9

Specification

• Models are described in languages

• Languages should be capable of representing the following
features:
• Hierarchies
- Behavioral hierarchies: containing objects necessary to describe the
system: States, events, and output signals are examples of such objects
- Structural hierarchies: how systems are composed of physical
components comprised of components: processors, memories, actuators,
and sensors
• Component-based design
• Concurrency
• Synchronization and communication
- It is necessary to express mutual exclusion
• Timing behavior
- we need to show that certain computations are completed in a given
amount of time, but the algorithm as a whole should possibly run until
power is turned off
• elapsed time, delaying of processes, timeouts, Deadlines

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 10

Specification

• Timing behavior
• we need to show that certain computations are completed in a given
amount of time, but the algorithm as a whole should possibly run
until power is turned off
- elapsed time, delaying of processes, timeouts, Deadlines

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 11

Specification (Cont.)

• Models are described in languages

• Languages should be capable of representing the following
features:
• State-oriented behavior
- State-oriented behavior provided by automata should be easy to
describe
- classical automata models are insufficient, since they cannot model
timing and since hierarchy is not supported
• Event-handling
• Exception-oriented behavior
• Presence of programming elements
• Executability
• Support for the design of large systems
• Domain-specific support
• Readability
• Portability and flexibility
• Termination
• Support for non-standard I/O devices

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 12

Specification (Cont.)

• Models are described in languages

• Languages should be capable of representing the following
features:
• Non-functional properties
• Support for the design of dependable systems
• No obstacles to the generation of efficient implementations
• Appropriate model of computation (MoC)

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 13

Model of Computation

• Here will not be any single formal language meeting all

these requirements
• In practice, we must live with compromises and possibly also with a mixture of
languages (each of which would be appropriate for describing a certain type of
problems)
• The choice of the language used for an actual design will depend on the
application domain and the environment in which the design has to be
performed

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 14

Model of Computation

• MoCs describe the mechanism assumed for performing

Computations
• Components and the organization of computations in
components: Procedures, processes, functions, and finite
state machines are possible components
• Communication protocols describe methods for
communication between components. Asynchronous message
passing and rendezvous-based communication are examples of
communication protocols

• Nodes in the graph represent components performing computations

• Their causal dependence is the most obvious relation between
computations

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 15

Model of Communication

• Models of communication:
• Shared memory
• Critical sections: include semaphores, mutexes, conditional critical
regions, monitors, and spin locks
• Shared memory-based communication can be fast but is difficult to
implement in multiprocessor systems without a common physical
memory

• Message passing
• Messages are sent and received
• It can be implemented easily even if no common memory is
available
• Asynchronous message passing, also called non-blocking communication
• Synchronous message passing or blocking communication,
rendezvousbased communication
• Extended rendezvous, remote invocation

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 16

Organization of computations within the components

• Von Neumann model: This model is based on the sequential

execution of sequences of primitive computations
• Differential equations: Differential equations are capable of
modeling analog circuits and physical systems
• Discrete event model: In this model, there are events carrying a
totally ordered time stamp, indicating the time at which the event
occurs
• A global event queue sorted by time
• Entries from this queue are processed according to this order
• Finite state machines (FSMs): This model is based on
the notion of a finite set of states, inputs, outputs, and transitions
between states. Several of these machines may need to
communicate, forming so-called communicating finite state
machines (CFSMs)
• Data flow: In the data-flow model, the availability of data triggers
the possible execution of operations
Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 17
Organization of computations within the components

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 18

Von Neumann vs Non-von Neumann
• Von Neumann model of sequential execution combined with
some communication technique is a commonly used MoC
• In this model, specifications will typically consist of tasks,
processes, or threads
• Von Neumann computing is implicitly based on accesses to
globally shared
memory → making deadlocks
• Rarely deadlocks may be acceptable for a PC, but they are
clearly unacceptable for a safety-critical system
• It makes sense to study non-von Neumann MoCs avoiding this
problem
• Non-von Neumann models are frequently called model-based
designs
• The key idea of model-based design is to have some abstract
model of the system under design (SUD)
Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 19
Organization of computations within the components

• Von Neumann model: This model is based on the sequential

execution of sequences of primitive computations
• Differential equations: Differential equations are capable of
modeling analog circuits and physical systems
• Discrete event model: In this model, there are events carrying a
totally ordered time stamp, indicating the time at which the event
occurs
• A global event queue sorted by time
• Entries from this queue are processed according to this order
• Finite state machines (FSMs): This model is based on
the notion of a finite set of states, inputs, outputs, and transitions
between states. Several of these machines may need to
communicate, forming so-called communicating finite state
machines (CFSMs)
• Data flow: In the data-flow model, the availability of data triggers
the possible execution of operations
Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 20
Modeling Continuous Dynamics- Differential Equations

• Ordinary differential equations, Laplace transforms,

feedback control models, …

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 21

An Example: Helicopter Dynamics

• Six degrees of freedom (6DoF):

• Three of these represent position in three
dimensional space: x, y, z
• Three represent orientation in space: pitch, yaw,
roll

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 22

Notation: Continuous-Time Signals

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 23

Notation: Differential Equation

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 24

Integral Equation

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 25

Integral Equation

• You can determine the acceleration, velocity, and

position of the object at any time
Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 26
Angular Version of Force: Torque
For a point mass rotating around a fixed axis:

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 27

Angular Version of Force: Torque
For a point mass rotating around a fixed axis:

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 28

Rotational Version of Newton’s Second Law

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 29

Feedback Control Problem

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 30

Plant and Controller

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 31

Simplified Model

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 32

Actor Models

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 33

Actor Models

• Actor: are waiting for input and perform their

operation once all required inputs have arrived
• Actor: A box like that above, where the inputs
are functions and the outputs are functions, is
called an actor

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 34

Actor Model of Helicopter

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 35

Cascade Composition

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 36

Actor Model with Multiple Inputs

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 37

Actor

• Closely related to the concept of function in

mathematics
• Can be used to model sensors, actuators,
computation, and plant

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 38

Key Point in Modeling a Plant

• The most important point in modelling a plant as

an actor is to correctly determine the input(s)
and output(s)

– Output is what you desire to control

– Input is what you really give (apply) to the system
from the viewpoint of the controller

Embedded and Real-Time Systems, Spring 2024, SBU, Tehran, Iran 39

Other Examples

• Actor model of the plant in the air conditioning

system

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 40

Other Examples

• Actor model of the plant in one dimensional

velocity control

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 41

Simulink Library

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 42

Simulink Library

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 43

 Feedback Control

Actual behavior

Desired behavior compensate to reduce future error

• Ψ: continuous-time system specifying the desired

angular velocity
• e: represents the difference between the actual and
the desired angular velocity

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 44

Feedback Control

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 45

Feedback Control

If we assume:

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 46

• End of Differential equations
• Next Lecture
– Other types of MoC

Embedded and Real-Time Systems, Spring 2025, SBU, Tehran, Iran 47