CYBER PHYSICAL SYSTEM HARDWARE PLATFORM

A Cyber-Physical System (CPS) Hardware Platform is the physical and computational

infrastructure that enables real-world cyber-physical systems to function. It includes processors, memory,
sensors, communication modules, and programmable logic (like FPGAs) integrated into a single platform.
CPS platforms connect computational algorithms (cyber part) with physical components
(sensors, actuators, etc.) to enable intelligent, real-time interaction with the environment.

Why do we need it?

 Traditional CPS simulators only simulate parts of the system, focusing mainly on performance.
 Existing tools are either:
o Accurate but slow, requiring heavy computation.
o Fast but less accurate, using high-level abstraction.
 There's a lack of tools that support full system design and analysis while being both accurate and
efficient.

1. COSSIM Framework
 COSSIM stands for Co-simulation and Simulation platform for CPS.
 It is an open-source framework

Purpose:
 To simulate both computing and communication components of CPS together.
 Provides high simulation speed with good accuracy, which is missing in most tools.
Use Case:
 Successfully tested in a visual search application (detects objects in images using a client-server
model).
 Also tested on an embedded system developed by STMicroelectronics.
2. AXIOM Project
 AXIOM = Agile, eXtensible, fast I/O Module
 It is a hardware-software platform for next-generation CPS.
Key Goals:
1. Support parallel programming with ease.
2. Scale performance by connecting multiple boards (1 to 10 or more).
3. Use FPGAs (Field Programmable Gate Arrays) to accelerate specific tasks.
4. Reduce system cost with efficient board interconnection.
Applications:
 Designed for smart surveillance and smart home living systems.

3. OmpSs Programming Model (used in AXIOM)

 A task-based parallel programming model.
 Supports coding in:
o OpenCL
o CUDA
o C/C++
 Converts code to run on:
o GPUs, CPUs, or FPGAs.
Features:
 Automatically distributes tasks across multiple processing nodes.
 Handles communication within distributed systems efficiently.
 Useful in designing heterogeneous systems (with different types of processors).
4. AXIOM Hardware Platform
 Combines ARM multi-core processor with FPGA fabric.
Scalability & Modularity:
 Can connect multiple AXIOM boards using:
o USB-C based AXIOM-Link (cheap and fast).
o 2 or 4 bi-directional links per board.
 Boards can be connected in:
o Ring topology
o 2D Mesh/Torus topology
 Allows for cost-effective, upgradable, and high-speed CPS systems.

5. AXIOM Network Interface Architecture

 Supports high-speed communication using:
o Remote Direct Memory Access (RDMA)
o Remote Write Operations
 These features are accessible at the application level, meaning software can directly manage
efficient memory sharing between different boards.

PROCESSORS, SENSORS, ACTUATORS, CPS NETWORK

They are the core components of Cyber-Physical Systems (CPS) that interact with the physical world.
These include:
 Processors that handle computation and control logic,
 Sensors that gather data from the physical environment,
 Actuators that perform actions based on decisions made by the system,
 and a Network that connects all these components and enables communication.
 These elements form a feedback loop: sensors detect physical conditions → data is processed →
decisions are made → actuators perform actions → and the cycle repeats.
Why is it needed?
 To build intelligent systems that observe, decide, and act in real time.
 Essential for applications like autonomous vehicles, smart grids, industrial automation, remote
health monitoring, and robotics.
 Helps systems to respond to environmental changes autonomously with minimal human intervention.
 Enables seamless coordination between distributed components through networking.

PROCESSORS
Processors are the computational core of CPS. They receive input from sensors, analyze the data,
and send control signals to actuators.
In communication-based systems, special processors are optimized for handling communication
protocols, real-time data transfer, and coordination among components.

Communication Processors
These processors are tailored to handle wired and wireless communication efficiently.
 Categories based on communication system type:
o Wired or Wireless processors
 Categories based on communication layer:
o Physical Layer Processors: handle modulation/demodulation
o MAC Layer Processors: manage access to shared communication channels
o Network Layer Processors: manage data routing, addressing, etc.
 Application-specific categorization:
o Audio, video, data processors
 End-system-based categorization:
o Laptops, mobile phones, embedded systems

Network Processors in Wireless Operation

In wireless CPS environments, specific processors are used to manage wireless communication efficiently.
Examples:
 SNAP Processor: Used for mesh networking and low-power wireless control.
 Subliminal Processor: Another example of a processor optimized for secure and stealthy
communication in wireless environments.
These processors coordinate wireless communication among CPS components like sensor nodes and
actuators.
SENSORS
Sensors are devices that detect physical conditions (like temperature, pressure, motion) and convert
them into electrical signals to be analyzed by a processor.

Importance
 Form the input mechanism of CPS.
 Monitor system behavior and environmental changes.
 Enable automation and fault detection (e.g., motor overheating).
 Provide data for decision-making and control.

Working
 Sense a physical change.
 Generate a signal (analog or digital).
 Send it to a processor or controller.
 The system responds (e.g., shutdown a machine, adjust speed, etc.).

Types of Sensors:
 Temperature Sensors: Detect temperature changes.
 Vibration Sensors: Monitor mechanical movements and vibrations.
 Security Sensors: Used in access control and surveillance.
 Pressure Sensors: Measure pressure in fluids or gases.
 Humidity Sensors: Track moisture levels in the environment.
 Gas Sensors: Detect the presence of gases like CO₂, methane, etc.
These are essential for environmental monitoring, safety, maintenance, and process optimization.

ACTUATORS
Actuators are devices that convert electrical signals into physical motion or force. They are the
output units in CPS that perform real-world actions based on decisions.

Importance
 Close the loop in CPS by acting on the physical world.
 Without actuators, CPS would only be able to monitor, not act.
  Used in robotics, automation, aerospace systems, and smart manufacturing.

Working
 Receive control signals from the processor.
 Use an energy source (electricity, air, hydraulics).
 Perform physical movement (e.g., rotate a shaft, open a valve).

Types of Actuators:
1. Manual Actuators:
o Operated manually (e.g., gears, levers, wheels).
o Simple, low-cost, but not suitable for automation.
2. Pneumatic Actuators:
o Use compressed air to generate motion.
o Common in industrial valve control.
3. Hydraulic Actuators:
o Use pressurized liquid.
o Can generate large forces, used in heavy machinery.
4. Electric Actuators:
o Use electric motors.
o Precise and efficient, but depend on power availability.
5. Spring Actuators:
o Hold mechanical energy in a compressed spring.
o Triggered by an event, releasing the spring for a one-time emergency action.
WIRELESS SENSOR NETWORKS (WSNS)
A WSN is a network of distributed sensor nodes that collect data and transmit it wirelessly to a
central processor or base station.
Importance
 Essential for scalable, flexible, real-time monitoring in CPS.
 Reduced need for physical wiring.
 Enables remote data collection in hard-to-access locations.

Technological Background
 Enabled by MEMS (Micro-Electro-Mechanical Systems) that allow miniature sensors with low
power consumption.
 Energy-efficient designs include solar, vibration, and wind energy harvesting.
 Can operate in infrastructure-based (planned) or unstructured (ad hoc) deployments.
Types of WSNs:
 Terrestrial WSN: Ground-level sensor networks.
 Underwater WSN: Sensors deployed underwater using acoustic communication.
 Underground WSN: Buried networks for soil/earth monitoring.
 Multimedia WSN: Capture images/audio.
 Mobile WSN: Nodes are mobile and adapt to changing environments.

ROLE OF SENSORS AND ACTUATORS IN CPS

  Sensors and actuators form the interface between the physical world and the cyber system.
 Multiple sensors collect environmental and performance data.
 This data is processed in real-time in the cyber system.
 The system sends control commands to actuators.
 Actuators carry out physical operations (adjust speeds, open valves, etc.).
Bidirectional Interaction:
 Changes in the physical world → affect the cyber world.
 Updates or commands in the cyber system → cause changes in the physical world.
This continuous feedback loop is the core principle of CPS.

NETWORKED CYBER-PHYSICAL SYSTEMS

Networked CPS are systems where cyber and physical components are connected via
communication networks, often wireless.

Importance:
 Enable real-time coordination between devices.
 Allow remote control, monitoring, and data analysis.
 Support cloud integration and big data analytics.

Challenges:
 Localization & positioning (especially indoors)
 Reliable wireless communication
 Security & privacy of sensitive data
 Scalability for large deployments
 Real-time responsiveness even in dynamic conditions

Participants in CPS Networks:

 Smart sensors
 Intelligent actuators
 Communication gateways
 Cloud servers
 Human-machine interfaces
 Mobile & wearable devices
These all must work in sync to create intelligent, adaptive, and secure systems.
 WIRELESSHART
As process industries expand or modernize, the demand for additional measurements rises—for
improved control, regulatory compliance, safety, and efficiency. However, running new wires for each
sensor is expensive and time-consuming. WirelessHART offers a secure, reliable, and cost-effective
solution to gather more data without laying cables.
 WirelessHART is a wireless extension of the well-established HART protocol (Highway
Addressable Remote Transducer).

 It is an open, interoperable standard for wireless communication between field instruments and
control/monitoring systems in industrial automation.
 Developed to support multi-vendor devices, it builds on HART’s vast installed base (~30 million
devices globally).

TECHNICAL ARCHITECTURE
Mesh Networking Features
 Time-Synchronized Mesh Protocol (TSMP):
o Based on Dust Networks' technology.
o Self-organizing and self-healing network.
 Devices work as routers, forwarding data from neighbors to expand range and improve redundancy.
 Uses IEEE 802.15.4 radios in the 2.4 GHz ISM band (common with ZigBee, Thread).
 Utilizes TDMA (Time Division Multiple Access) and channel hopping, a technique used in wireless
networks to reduce interference and improve communication.

Communication Reliability
 Direct Sequence Spread Spectrum (DSSS) for secure communication.
  Channel Hopping ensures robustness in noisy RF environments.
 If a route fails, packets auto-reroute using alternative paths (similar to internet routing).
 Continuous evaluation of paths based on latency, efficiency, and reliability.
CORE COMPONENTS OF WIRELESSHART NETWORK
1. Field Devices
 WirelessHART-enabled transmitters (or HART devices with adapters).
 Collect process data like temperature, pressure, flow, etc.
 Each device may act as a router for others in the mesh.
2. WirelessHART Gateway
 Central point of the network.
 Translates WirelessHART data to host protocols (Modbus, Profibus, Ethernet).
 Typically includes:
o Network Manager (controls routing and scheduling),
o Security Manager (manages encryption and authentication).
3. Network Manager
 Responsible for:
o Slot allocation (WirelessHART splits each second into 100 × 10ms time slots),
o Network topology,
o Communication scheduling and
o Link health monitoring.
4. Security Manager
 Distributes and manages encryption keys.
 Ensures only authorized devices can join.
 Provides confidentiality and authentication using AES-128 encryption.
Other Supporting Components
🔹 Repeater
 A node that only routes traffic, with no process connection.
 Useful for bypassing obstacles or extending network range.
🔹 Adapter
 Bridges existing 4–20 mA HART devices to the wireless network.
 Can be:
o Battery-powered,
o Powered via the 4–20 mA loop,
o Or even power the instrument itself.
 Some disable analog output completely and rely entirely on WirelessHART.
🔹 Handheld Terminal
 Two versions:
1. Standard HART FSK: for regular device configuration using Device Description (DD) files.
2. Wireless-enabled handheld: communicates with instruments via the WirelessHART mesh
through the gateway—for diagnostics, joining networks, etc.

MESSAGE TYPES SUPPORTED

WirelessHART is highly flexible:
 Periodic publishing of measurement values.
 Event-driven messages (e.g., alarm conditions).
 Request/response commands.
  Large data transfers with segmentation support.
This allows communication patterns to be optimized for power efficiency and application needs.

ADVANTAGES OF WIRELESSHART
 Backward Compatibility: Works with existing HART infrastructure.
 Cost Efficiency: Saves cost by eliminating wiring.
 High Reliability: Redundant routing and healing mesh ensure data gets through.
 Security Built-In: AES encryption and device authentication.
 Scalability: New devices can be added with minimal planning.
 Power Efficient: Optimized for battery-powered devices with low overhead.

CAN IN AUTOMOTIVE SYSTEMS

 CAN stands for Controller Area Network.
 Developed specifically for the automotive industry to enable efficient and reliable communication
between various Electronic Control Units (ECUs) in a vehicle.
 Replaces complex point-to-point wiring with a common serial bus, reducing wiring complexity and
cost.

Features
 Multi-master, message-based protocol.
 Each node (device/ECU) can send and receive messages.
 Uses unique message IDs to filter and prioritize communication.
 Arbitration ensures that the highest priority message wins when multiple nodes transmit
simultaneously.

CAN BUS ARCHITECTURE

 Consists of three main components per node:
o Microcontroller Unit (MCU) – handles main logic and processing.
o CAN Controller – formats data into CAN messages (may be integrated into MCU).
o CAN Transceiver – converts data from the controller to differential signals (CANH/CANL)
and vice versa.
Electrical Characteristics
 Differential signaling with two wires: CAN High (CANH) and CAN Low (CANL).
 Idle voltage: both at ~2.5V.
 Dominant bit (logic 0): CANH = 3.5V, CANL = 1.5V → Differential = 2V.
 Recessive bit (logic 1): both = 2.5V → Differential = 0V.
 120-ohm termination resistors at each end of the bus to prevent signal reflection.
CAN LAYERED ARCHITECTURE
 Application Layer – Interfaces with applications or the OS.
 Data Link Layer – Handles framing, error checking, and arbitration.
 Physical Layer – Consists of the transceiver and physical wiring.

CAN BUS IN DEVELOPMENT BOARDS

Arduino
 Uses CAN shield with external CAN controller (e.g., MCP2515) and transceiver (e.g.,
MCP2551).
 CAN communication via SPI interface.
Raspberry Pi
  No native CAN controller.
 Requires external CAN controller (e.g., MCP2515) via SPI.
 Connected to a CAN transceiver (e.g., TJA1040, MCP2551).
STM32 Microcontrollers
 Native support for CAN1 and CAN2 with built-in CAN controller.
 Requires only an external transceiver for complete CAN functionality.

ADVANTAGES OF CAN
 Robust communication with high fault tolerance.
 Reduced wiring complexity.
 Scalable – new ECUs can be added easily.
 Real-time performance due to deterministic arbitration.
 Widely supported by automotive-grade microcontrollers and development boards.

AUTOMOTIVE ETHERNET
 Automotive Ethernet is a specialized version of traditional Ethernet adapted for use in vehicles.
 Unlike standard Ethernet (which uses separate wires for transmit and receive), automotive Ethernet
uses a single twisted pair to perform bidirectional communication—transmitting and receiving data
simultaneously.

 Initially used for diagnostics, infotainment, and sensors, it is now a strong candidate for autonomous
vehicle systems due to its high bandwidth and scalability.

Why Use Ethernet in Vehicles?

  Automotive systems are evolving due to the push toward automation, connectivity, and data-
intensive applications (e.g., cameras, radar, LiDAR).
 Legacy networks like CAN, LIN, and FlexRay offer limited bandwidth (kbps to low Mbps), which
cannot keep up with the demands of modern automotive data systems.
 Ethernet, a well-established and widely-used networking standard, is now being adopted due to its
performance, scalability, and cost-effectiveness.

BENEFITS OF AUTOMOTIVE ETHERNET

1. High Bandwidth
 Supports 100 Mbps today and is scaling to multi-gigabit speeds
 Ideal for applications like:
o Advanced Driver Assistance Systems (ADAS)
o Vehicle-to-Everything (V2X) communications
o Surround-view cameras and radar/LiDAR systems
2. Lightweight and Cost-Effective
 Uses single unshielded twisted-pair (UTP) cabling.
 Reduces:
o Cable weight by ~30%
o Connectivity cost by ~80%
3. Scalability and Flexibility
 Switch-based Ethernet topology allows for easy expansion and structured networking.
 Can support multiple ECUs, sensors, and systems without complex rewiring.
4. Compatibility with IT infrastructure
 Ethernet enables easier integration with cloud platforms, diagnostics, and smart traffic systems.
 Common with smart infrastructure, facilitating V2X (Vehicle-to-Everything) communication.

AUTOMOTIVE ETHERNET VS. OTHER PROTOCOLS

Protocol Max Speed Usage Area Drawbacks

CAN Up to 1 Mbps Control systems Low bandwidth, not suitable for AV
LIN Up to 20 Kbps Body electronics Very low speed
FlexRay Up to 10 Mbps Safety-critical systems Higher complexity and cost
MOST 100–150 Mbps Infotainment systems Proprietary, heavier, fragile fibers
Ethernet 100 Mbps → Multi-Gbps Infotainment, ADAS, AV Needs robust security implementation

CHALLENGES
  Cybersecurity risks: Increased connectivity and compatibility with external networks make
Ethernet systems more vulnerable to attacks.
 Real-time performance: While Ethernet supports high speed, strict real-time determinism (needed
in critical control systems) still requires additional protocols or enhancements.

REAL-TIME SCHEDULING IN CYBER-PHYSICAL SYSTEMS (CPS)

In real-time systems like robot control, avionics, and multimedia processing, timing is critical. With
CPS—where computational (cyber) and physical processes interact closely—spatial (location/movement)
factors must be considered alongside traditional timing factors.

Difference between Conventional Real-Time scheduling and Real-Time scheduling for CPS
Conventional Real-Time
Functions Real-Time scheduling for CPS
scheduling
Scheduling resources CPU, BW, memory, I/O Servicing node
Scheduling environment Cyber environment Cyber and physical environment
Cyber factors (Period, Cyber factors and physical factors (period,
Scheduling parameters execution time, release time, execution time, release time, deadline,
deadline, etc.) migration delay time, etc.)
No migration is required for
Migration Migration is required for CPU and Job
CPU and Job.
Well-known Scheduling
RM, EDF, LST No standard dominant algorithm yet
algorithm
Do not consider spatial Physical migration delay time (between
Spatial issues
issues servicing node to serviced node)
Execution time, release time, effective
Execution time, release
Considering issues deadline (deadline—moving time), effective
time, deadline, laxity
laxity (laxity—moving delay time)
Scheduling Model Assumptions for FPS (Fixed Priority Scheduling)
 System has a uniprocessor, using preemptive fixed-priority scheduling.
 Tasks are periodic and schedulable.
 Slack time (li): exponential distribution.
 Execution time (ei): uniform distribution.
 Migration time (mi): time to move computing node to where the task needs to be serviced.
 Cloud servers may be available (e.g., AWS), but not always predictable or reliable in timing.

LST vs ELST vs O-ELST vs H-ELST

Algorithm Key Idea

LST (Least Slack Time) Prioritize task with least slack time (laxity = deadline - release - exec time)
LST + includes physical movement delay (effective laxity = laxity - migration
ELST (Effective LST)
delay)
O-ELST (Optimal
Improves ELST by reordering tasks only if schedule improves performance
ELST)
Balances slack and movement using a weighted sum:
H-ELST (Heuristic
priority = α × slack time + (1−α) × migration delay
ELST)
Lower values = higher priority. α determines importance of time vs. space
Important Considerations in CPS Scheduling
 Effective Deadline: deadline - migration delay
 Effective Laxity: laxity - migration delay
 Tradeoff Tuning: Use α (between 0 and 1) to balance cyber (timing) vs. physical (movement)
constraints.
 Heuristics like H-ELST are more scalable for real systems with lower computational overhead than
optimal ones.

DYNAMICAL SYSTEMS
A dynamical system is any system whose behavior evolves over time, typically governed by physical
laws. In control systems, we design controllers that interact with the physical world via:
 Sensors – to measure state variables (e.g., temperature, speed).
 Actuators – to influence the system (e.g., heating system, throttle).

Control System Structure

 A control system manages physical processes via sensors (measurements) and actuators
(commands).
 The physical system being controlled is called the plant.
 The controller regulates the plant based on:
 o Reference inputs (desired values),
o Sensor measurements (current state),
o Disturbances (external, uncontrolled factors like road slope).

Continuous-Time Dynamical Systems

 These systems have inputs and outputs that evolve continuously over time (not in discrete steps).
 Example variables: temperature, velocity, position.
 Variables are often real numbers within specific ranges (e.g., v ∈ [0, 150] mph).

Mathematical Modeling
 A typical continuous-time system has:
o State variables (e.g., position x, velocity v)
o Input variables (e.g., force F, road angle θ)
o Differential equations describing their evolution.
Example: Car on a Graded Road
 Motion model:
dx/dt = v
dv/dt = (F - kv - mgsin(θ)) / m
where:
o F = engine force (controlled),
o kv = friction,
o mgsin(θ) = gravitational effect (disturbance),
o θ = slope angle (uncontrolled disturbance).
Disturbances
 Variables like θ (road grade) are not controlled by the system.
 Controllers must account for disturbances to maintain stability and performance.
 Must ensure robustness: system performs well even when disturbances vary.

Component Composition
 Components can be connected via block diagrams (similar to synchronous components).
 Ensure:
o No cyclic await dependencies,
o Well-defined inputs and outputs.
 Output variable y awaits input x if y(t) depends on x(t).

STABILITY IN DYNAMICAL SYSTEMS

In control systems (like cruise controllers or pendulums), stability ensures that small disturbances
(like a small bump or slope on the road) do not lead to large, uncontrolled changes in behavior (like
vehicle speed). This is especially important for systems that evolve continuously over time.

Safety and Liveness vs. Stability

 Safety: Something bad never happens (e.g., speed always below max).
 Liveness: Something good eventually happens (e.g., car reaches desired speed).
 Stability (new, specific to continuous systems): Small input disturbances should not cause large
output changes.

EQUILIBRIUM STATE
An equilibrium is a steady state where the system doesn’t change over time unless disturbed.
 Example: If a car is parked (velocity = 0), and there's no force applied, it stays still. That’s an
equilibrium.
Different systems can have multiple equilibria:
 A car at any position but with 0 speed is in equilibrium.
 A pendulum hanging straight down or straight up (perfectly balanced) are both equilibria, but they
behave differently when disturbed.

LYAPUNOV STABILITY
This refers to how a system reacts when slightly disturbed from its equilibrium.
 A system is stable if it stays close to the equilibrium after a small disturbance.
 It is asymptotically stable if, after the disturbance, it not only stays close but also eventually
returns to the equilibrium state over time.
Examples:
 The car slowing down to rest (but not returning to the exact starting position) is stable but not
asymptotically stable.
 A pendulum swinging back and forth without stopping is also stable but not asymptotically stable.
 If the pendulum had friction (not shown in the model), it would eventually come to rest — that
would be asymptotic stability.

BOUNDED-INPUT BOUNDED-OUTPUT (BIBO) STABILITY

This is a different perspective — it looks at the system as a black box:
 If you feed a signal (input) into the system, and the input is not extreme (i.e., it’s bounded),
 Then the output should also stay within a reasonable (bounded) range.
Example:
 A helicopter rotor getting a constant torque will keep spinning faster and faster, even though the
torque doesn’t change. This is not BIBO stable — a small, constant input leads to an unbounded
(growing) output.

DESIGNING CONTROLLERS
In control systems, a controller is designed to provide inputs to a dynamical system (plant) so that
its output follows a desired target, even when external disturbances are present.

OPEN-LOOP VS FEEDBACK CONTROLLERS

Open-Loop Controller
 It does not use sensor data to make decisions.
 It relies solely on the mathematical model of the system.
 Example: A car maintaining constant speed by applying a fixed force (assuming conditions don't
change).
  Cheap and simple, but can fail if the environment or system changes.
 Works only when the system is predictable or can be manually adjusted.
Feedback Controller
 Uses sensors to measure output or state and adjusts input accordingly.
 Example: If a car slows down going uphill, the controller senses the speed drop and increases the
force.
 Can handle unpredictable disturbances and works even if the model is not perfect.
 Requires sensors, and its effectiveness depends on their accuracy.

STABILIZING CONTROLLER
 Some systems (like a helicopter's spin) are inherently unstable.
 A proportional controller is used: It calculates how far off the system is from the desired value
(error) and applies a correction proportional to that error.
 The controller aims to reduce the error and stabilize the system over time.
 If the error is large, the correction is large; if small, the correction is small.
 Increasing the correction factor (gain) makes the system stabilize faster — but it must still be within
safe limits.

PID CONTROLLERS
PID controllers are widely used in industrial control systems to make sure that the output of a system
(like the speed of a motor) follows a desired reference (like a target speed). They do this by adjusting the
system's input based on:
1. Proportional (P): How far off the output is from the target right now.
2. Integral (I): How much total error has built up over time.
3. Derivative (D): How quickly the error is changing.

DC Motor Control Example

A DC motor converts electrical energy (input voltage) into mechanical rotation. It’s a good example for
studying PID control because it's common in many devices.
 When voltage is applied to the motor, it causes current to flow and the motor shaft to rotate.
 The motor’s behavior is affected by electrical resistance, inductance, rotational inertia, and friction.
  As the shaft rotates, it generates a voltage that opposes the input voltage (called back EMF), making
the system self-regulating to an extent.

Controlling the Motor

The goal of the controller is to adjust the input voltage so that the rotational speed of the motor reaches
and stays at a desired value.
Proportional Controller (P-only)
 Only reacts to current error.
 Speeds up response but doesn't eliminate long-term errors (called steady-state error).
 Increasing the proportional gain improves responsiveness but may cause overshooting and
oscillations.

System Response Terms (based on output graph):

1. Rise Time: How fast the system reaches the target value initially.
2. Overshoot: How much the output exceeds the target before settling.
3. Steady-State Error: How far the final output is from the target.
4. Settling Time: How long it takes for the output to settle near the target without further oscillation.

PI and PD Controllers
 PI (Proportional + Integral): Helps remove steady-state error by considering cumulative past
errors, but can lead to more overshoot and longer settling time.
 PD (Proportional + Derivative): Helps reduce overshoot and smoothens the response by reacting to
how fast the error is changing.

Full PID Controller

 Combines all three components.
 Offers the best control performance: quick rise, minimal overshoot, no steady-state error, and fast
settling.
 Tuning the gains for P, I, and D is key to achieving optimal results.

PERFORMANCE UNDER PACKET DROP AND NOISE

In real-world dynamic systems—especially those involving networked control systems (NCS),
wireless communication, or distributed sensors/actuators—two major issues can significantly affect
performance: packet drop and noise.

PACKET DROP IN DYNAMIC SYSTEMS

 Packet drop refers to the loss of data packets during transmission between components of a control
system (e.g., between sensors, controllers, and actuators). This is common in systems using wireless
networks or the internet.

Effects on System Performance:

 Delayed or missed control signals: The system may not respond in time to disturbances or desired
changes.
 Instability: Frequent packet losses can destabilize a previously stable system.
 Increased settling time and overshoot: Due to irregular updates, control actions may be applied too
late or inappropriately.
 Poor tracking performance: In tracking tasks (e.g., following a reference signal), packet drop
causes deviation from the desired trajectory.

Common Mitigation Strategies:

 Buffering and acknowledgment mechanisms (at network layer).
 Predictive control: Estimating future system states to compensate for missing data.
 State estimators (like Kalman filters) that can handle missing measurements.
 Robust controllers or event-triggered control that can work with intermittent updates.

NOISE IN DYNAMIC SYSTEMS

Noise refers to unwanted disturbances that corrupt sensor measurements, control signals, or system
states. It can be due to electromagnetic interference, thermal effects, quantization error, etc.

Effects on System Performance:

 Reduced accuracy of measurements: Affects feedback quality and may result in incorrect control
actions.
 Fluctuating control effort: Controllers may overreact to noisy signals, increasing wear and energy
consumption.
 Degraded estimation: Observers like Kalman filters may give less reliable state estimates in the
presence of high noise.
 Oscillations or instability: Especially for high-gain controllers reacting to noisy signals (especially
derivative terms in PID).
Common Mitigation Strategies:
 Signal filtering: Low-pass filters, moving average, or Kalman filters to clean noisy signals.
 Controller tuning: Reducing derivative gain in PID or using PID variants (e.g., PI or filtered PD).
 Sensor fusion: Combining multiple sensor inputs to improve accuracy and reduce noise effects.
 Robust control techniques: e.g., H∞ control or sliding mode control designed to handle model
uncertainty and disturbances.