U3 Flashcards
1
Q
M2M involves:
A
an autonomous device communicating directly to another autonomous device. In this context, autonomous refers to the ability of a node to start the communication with a neighboring node without human intervention.
2
Q
In M2M, devices do ——— necessarily depend on an ———————–connection. In such cases, the devices may communicate over ——————— channels, e.g., a serial port or ————–protocol. A simple example of M2M is controlling electrical appliances, such as bulbs and fans, using ———————— from a smartphone.
A
not
Internet
non-IP-based
custom
RF or Bluetooth
3
Q
Alternatively, IoT may incorporate some M2M?
A
nodes, but aggregates data at an edge router or gateway connected to the internet.
4
Q
Other scenarios implement the internet networking capabilities on each IoT device to?
A
separately deliver its data to cloud services.
5
Q
In fact, several technological advancements have driven the evolution from M2M to IoT, including?
A
the cloud technologies advancement,
the pervasiveness of wireless and mobile communication, the cost-effective new energy devices like lithium-ion,
and the arrival of advanced deep learning and AI tools.
6
Q
In fact, several technological advancements have driven the evolution from M2M to IoT, including the cloud technologies advancement, the pervasiveness of wireless and mobile communication, the cost-effective new energy devices like lithium-ion, and the arrival of advanced deep learning and AI tools. In this manner, IoT is a broader term than M2M solutions as?
A
it encompasses much more technology and connectivity.
7
Q
DevOps engineer:
A
These engineers will work with IT developers to facilitate better coordination among operations, development, and testing functions by automating and streamlining the integration and deployment processes.
8
Q
In general, IoT ecosystems start with sensors deployed in a certain location to ?
A
convert physical phenomena, such as movement, temperature, and pressure, into digital signals.
Such signals represent the data to be transmitted to the Internet.
9
Q
At this stage, the collected data traverse various ———–arriving at a ——————–. In fact, the powerful potential of IoT emerges from collecting millions of ———————–. To build such an ecosystem, we need experts from different ————————–disciplines.
A
channels
cloud service
sensory data
engineering
10
Q
At the outset, physicists are required to?
A
develop new sensor technologies and long lifetime batteries.
11
Q
Embedded system engineers are necessary to?
A
drive the sensors at the edge.
12
Q
For data collection, we need:
A
=network engineers who are capable working in a personal area network (PAN)
= or wide area network (WAN)
=as well as a software-defined networking.
=Furthermore, data scientists are needed to develop novel data analysis and machine learning algorithms at the edge and at the cloud.
= Finally, DevOps engineers have the responsibility of deploying scalable cloud solutions in addition to fog solutions.
13
Q
Obviously, IoT architecture may comprise several technologies and connectivity solutions. Hence, there exist a myriad of design choices for cloud storage:
A
IoT security systems,
networking,
and data analytics (Lea, 2018).
14
Q
For example,
A
selecting the wrong PAN protocol may lead to poor communication and significantly low signal quality.
15
Q
Selecting the wrong PAN protocol may lead to? interference effects in the local area network (LAN) and WAN and how costly the loss of data is. Furthermore, a decision must be made about which Internet protocols, such as message queuing telemetry transport (MQTT), constrained application protocol (CoAP), and advanced message queuing protocol (AMQP), to adopt. In the context of data processing, a system designer must decide whether or not to apply fog computing via processing data close to its source to solve latency problems and to reduce bandwidth and communication costs.
A
poor communication and significantly low signal quality.
16
Q
System designers have also to consider?
A
interference effects in the local area network (LAN)
and WAN and how costly the loss of data is.
17
Q
Furthermore, a decision must be made about which Internet protocols, such as?
A
message queuing telemetry transport (MQTT), constrained application protocol (CoAP), and advanced message queuing protocol (AMQP), to adopt.
18
Q
In the context of data processing, a system designer must decide?
A
-whether or not to apply fog computing via processing data close to its source to solve latency problems
- and to reduce bandwidth and communication costs.
19
Q
The figure below demonstrates the main components of IoT system architecture with several design options. Specifically, there exist five main components:
A
(1) sensing and power;
(2) data communication;
(3) Internet and routing protocols;
(4) cloud and fog computing;
and (5) IoT security.
20
Q
The first component involves?
A
any device that is capable sensing the world.
In many cases, a single sensor can generate a massive amount of data,
such as auditory sensing for preventative maintenance of machinery.
Other applications may only acquire a single bit of data indicating, for example, vital health data from a patient.
21
Q
In general, sensors have widely grown in scale up to ————————- sizes with significant cost ————. To drive the sensors at the edge,————————————- power systems are required. Collections of billions of small sensors still require a ———————- amount of energy to power. Therefore, novel energy supply methods, such as ————–, have been recently developed to enable sensors to function for many years.
moving the sensory data from the edge to cloud services and data centers located, for example, at Google, Amazon, Microsoft, and IBM. Such sensors are mostly battery powered, with minimal computing and storage resources.
A
sub-nanometer
reduction
low-size and low-cost
massive
harvesting
22
Q
The second component deals with?
A
moving the sensory data from the edge to cloud services and data centers located,
for example, at Google, Amazon, Microsoft, and IBM. Such sensors are mostly battery powered, with minimal computing and storage resources.
23
Q
Owing to these resources’ constraints, there exist several communication challenges :
A
-addressing and identification. According to the concept of IoT, billions of devices are to be connected to the Internet. Accordingly, each device has to be identified through a unique address. Hence, we need a large address space and a unique address for each IoT device.
low power communication. Wireless communication typically consumes a significant amount of energy. Therefore, we need solutions that facilitate data communication with low power consumption.
routing protocols with low memory requirement and efficient communication patterns.
high-speed and non-lossy communication.
mobility of smart IoT devices.
24
Q
In addition to the IP-based communication (which is used by IoT devices to connect to the Internet), non-IP networks are?.
A
also used to locally connect the IoT devices
25
Non-IP communication channels, such as Bluetooth, RFID, and NFC, are popular but?
limited in their range (up to a few meters). Accordingly, these facilities are limited to small PANs widely used in IoT applications such as wearables connected to smartphones.
26
To design successful IoT systems, preliminary tools and models have to be?
incorporated, such as wireless radio dynamics like range and power analysis, signal-to-noise ratio, path loss, and interference.
27
The leading communication technologies used in the IoT world are?
IEEE 802.15.4,
low-power Wi-Fi,
6LoWPAN,
RFID, NFC,
Sigfox, LoRaWAN,
and other proprietary protocols for wireless networks.
28
The third component deals with?
bridging IoT data from sensors to the Internet through gateway routers
and supporting IP-based protocols.
Specifically, the primary role of the routers is securing, managing,
and steering the sensory data.
In fact, a single edge router can serve thousands of IoT devices.
29
IoT data needs :
efficient,
power-aware,
and low-latency protocols that can be easily steered and secured in and out of the cloud.
30
The fourth component deals with ?
processing the collected data by means of various aspects of cloud architectures such as SaaS, IaaS, and PaaS systems.
31
Nevertheless, not everything belongs in the ----------. There is a measurable cost in ------------------ data to a cloud vs. processing it at the ------------(edge computing) or extending cloud services ------------------------into an edge router (fog computing).
cloud
moving all IoT
edge
downward
32
Thanks to the advanced analytics and rules engines, the collected sensory data are mostly turned into?
=a set of decisions and actionable consequences. In some scenarios, a simple rules engine,
=e.g., fuzzy logic controller, can be adopted to easily detect anomalous temperature extremes on an edge router monitoring several IoT devices.
=In other scenarios, a massive amount of structured and unstructured data may be streaming in real time to a cloud service while requiring both fast processing for predictive analytics and long-range forecasting using advanced machine learning models.
33
Since the IoT devices are mostly in the public, in very remote areas, moving vehicles, or even inside a person, several security threats?
emerge. Several security primitives, such as asymmetric encryption, have been discussed.
34
In general, sensors are:
devices that can monitor and measure the physical world.
35
As the primary source of IoT data, ------------------represent the most important building block of all IoT applications. IoT sensors are mostly --------------------------------------------------------------------------------------. These IoT sensors come in a variety of forms and complexities, from simple -----------------------to advanced -----------systems.
sensors
small, low cost, and constrained by battery capacity
thermocouples
video
36
With the increasing popularity of smartphones, many smart IoT solutions have recently been built using smartphones because of ?
their embedded sensors.
37
Examples of the sensors embedded in modern smartphones are:
accelerometers,
gyroscopes,
cameras,
microphones,
GPS,
light sensors,
and proximity sensors.
38
In general, accelerometers sense?
the motion and acceleration through measuring changes in velocity of the smartphone in three dimensions.
39
Similarly, gyroscopes precisely detect the orientation of the smartphones through measuring?
capacitive changes when a seismic mass moves in a particular direction.
40
Modern accelerometers and gyroscopes are?
-fabricated as micro-electromechanical systems (MEMS).
-MEMS sensors incorporate miniaturized mechanical structures that can spin, stretch, bend, move, or alter form which in turn affects an electrical signal.
41
The figure below shows the components and basic principle of accelerometers
42
. Accelerometers use a MEMS piezoelectric to?
produce a voltage in response to movement.
Specifically, a central mass is attached to a spring which responds to acceleration in a certain direction.
43
When the accelerometer experiences an acceleration, the mass deflects from its ----------------------------. This deflection can be measured by varying ----------------------in a ----------------------------. Accelerometers are designed to respond to ------- dimensions (X, Y, Z) rather than --------------dimension.
neutral position
capacitance
MEMS circuit
multiple
one
44
Alternatively, gyroscopes operate slightly?
differently
45
A disk is driven to?
rotate a fraction of a full turn around its axis.
46
The tilt of the disk is measured to?
produce a signal related to the rate of rotation.
47
Both gyroscopes and accelerometers require?
power supplies and an op-amp for signal conditioning.
48
After conditioning, the output can be directly sampled by?
a digital controller.
49
These sensors can be manufactured in very small packages. For instance,
the Invensense MPU-6050 comprises a 6-axis gyro and accelerometer in a small 4 mm x 4 mm x 0.9 mm package with operating current of 3.8 mA, thus it is well-suited for low power sensing.
50
Additionally, magnetometers measure?
the strength and/or direction of magnetic fields.
51
Additionally, magnetometers measure the strength and/or direction of magnetic fields. This can be used as ?
a digital compass and in applications to detect the presence of metals.
52
Magnetometers are mostly harnessed to obtain?
directional information in three dimensions by being paired with accelerometers and gyroscopes.
53
An inertial measurement unit (IMU).
Magnetometers are mostly harnessed to obtain directional information in three dimensions by being paired with accelerometers and gyroscopes.
This device is called an inertial measurement unit (IMU)
IMUs are typically used to obtain location information in indoor environments.
54
To estimate the location information in outdoor environments?
-GPS sensors use a network of about 30 satellites orbiting the earth at an altitude of 20,000 km.
-The location is detected using the principle of trilateration in which the intersection point among three or more circles, whose centers are the satellites, is determined.
55
Aside from location detection, light sensors have recently been used to?
measure the intensity of ambient light.
56
Smart home applications employ such sensors to control ?
the lights in a room without human intervention. Furthermore, light sensors are used in many other IoT sensing activities, such as security systems, smart switches, and smart street lighting.
57
There exist two types of light sensors:
namely photoresistor and photodiodes.
58
Photoresistor varies in resistance depending on?
light intensity, while photodiodes convert light into an electrical current.
59
Similarly, proximity sensors employ infrared (IR) signals to measure?
the distance between the sensor and a certain object.
The main idea is to emit IR signals and wait for reflections from an object. Based on the difference in time, we can calculate the distance.
60
As an example,
proximity sensors can be used in applications in which we have to trigger some event when an object approaches the phone.
61
For long ranging and scanning, light detecting and ranging (LiDAR) sensors estimate the distance to an object by measuring?
a laser pulse reflection on the object.
62
The emitted laser power is typically constrained for safety reasons to prevent?.
eye damage
63
LiDAR sensors are heavily used in?
agriculture, automated and self-driving vehicles, robotics, surveillance, and environmental studies.
64
They are not only capable of estimating the range but are also capable of ?
analyzing anything that crosses its path.
65
For instance,
they can analyze gases, atmospheres, cloud formations and compositions, particulates, the speed of moving objects, and so on.
The figure below shows the utilization of LiDAR sensors in the automated sector to estimate the distance between moving vehicles.
66
In this realm,:
Google,
known for this project as Waymo (n.d.)
and Uber (2019) are developing self-driving vehicles.
Their vehicles feature a bulky box on top of the roof which spins continuously giving 360° visibility and precise, in-depth information about the exact distance to an object to an accuracy of ±2 cm
67
Some smartphones have a thermometer, barometer, and humidity sensor to measure ?
the temperature,
atmospheric pressure,
and humidity, respectively.
68
In IIoT, temperature sensors typically exist?
almost anywhere, e.g., smart thermostats, IoT cold storage logistics, refrigerators, and industrial machinery.
69
In general, there are three common temperature sensors in the IoT market:
thermocouples,
resistance temperature detectors (RTD), and thermistors.
70
Thermocouples produce very small signals, e.g., microvolts in amplitude?
because thermocouples do not receive an excitation signal to operate.
71
Thermocouples are well suited for?
- long distance measurements with long leads
- and are often used in industrial and high-temperature environments.
72
Thermocouples may suffer from aging? Hence,
i.e., high-heat environments can reduce the accuracy sensors over time.
73
IoT solutions must account for changes over the life of a sensor.
Thermocouples may suffer from aging
74
Alternatively, RTDs require an excitation current to operate?
They operate within a narrow range of temperatures but have much better accuracy than thermocouples.
75
RTDs' application in industry is limited?
Because RTDs are rarely used above 600˚C
76
Finally, thermistors are resistors that ?
change based on temperature and are suitable where a high resolution is needed for a narrow temperature range.
77
Thermistors are heavily used in:
medical devices,
scientific equipment,
food handling equipment,
incubators,
and home appliances such are thermostats.
The following table summarizes the main
78
differences between the three types of temperature sensors.
79
Comparison between the three types of temperature sensors
Temperature range
-----------------------------
Thermocouples -180 to 2,320
Resistance temperature detectors -200 to 500
Thermistors -90 to 130
Response time
-----------------------------
Thermocouples Fast (microseconds)
Resistance temperature detectors Slow (seconds)
Thermistors Slow (seconds)
Accuracy
-----------------------------
Thermocouples Low
Resistance temperature detectors Medium
Thermistors High
80
In general, wearable sensors are equipped with?
medical sensors that are capable measuring different parameters, e.g., the heart rate, pulse, blood pressure, body temperature, respiration rate, and blood glucose levels.
81
Such wearables comprise ---------------------------------------------------------------------------. In fact, smart watches and fitness trackers recently became popular as companies such as Apple, Samsung, and Sony (2016) are-------------------------- innovative features, such as connectivity with smartphones, sensors such as an --------------------------------.
smart watches, wristbands, monitoring patches, and smart textiles
incorporating
accelerometer, and heart rate monitors
82
In the same category, smart patches are pasted on ?
the skin to continuously monitor vital health parameters.
83
All the electronic components are?
embedded in these rubbery structures, and they are stretchable, disposable, and relatively cheap.
84
The figure below demonstrates a microneedle containing a smart insulin patch that is designed to sense elevated blood glucose levels.
85
It responds to abnormal glucose levels by?
releasing insulin. In fact, such smart patches offer people with diabetes a less-painful and more-reliable way to manage their condition.
86
RFID is?
an identification technology which consists of two main components: a tag and a reader.
87
The tag is?
a small chip with an antenna to periodically,
or upon interrogation,
transmit short,
digital,
radio frequency (RF) messages.
Such messages typically contain a unique identification code as well as some data stored in the tag’s memory.
88
To read the information encoded on a tag,?
the RFID reader emits a signal to the tag using an antenna.
89
The tag responds with the information written in its ----------------------- and the reader will then ----------------------the read results to an------------------ computer program.
memory
transmit
RFID
90
Furthermore, it measures the received signal strength (RSSI) of the received RF signal which is ?
an indicator of the range from tag to reader.
91
There exist two types of RFID tags:
active and passive.
92
Active tags have ?
a power source and passive tags do not have any power source.
93
Passive ?
tags draw power from the electromagnetic waves emitted by the reader and are thus cheap and have a long lifetime.
94
RFID sensors are employed in ?
many IoT applications, such as inventory management, supply chain management, asset tracking, access control, identity authentication, and object tracking.
95
For instance,
an RFID tag is attached to the front of vehicles. When the vehicle reaches a barricade on which there is a reader, it reads the tag data and decides whether it is an authorized vehicle. If the vehicle is found to be authorized, the barricade opens automatically.
96
In such an application, RFID cards are?
issued to the people who can then be identified by an RFID reader and given access accordingly.
97
Modern smartphones include:
powerful microphones and camera sensors for capturing audio and visual information, respectively.
98
Readings from both sensors can be?
=analyzed to detect various types of contextual information,
=such as the surrounding environment and users' activities.
= Furthermore, sound and vibration measurements are common in IIoT and predictive maintenance applications.
99
For instance,
an industrial machine that rotates or mixes a load of material in chemical manufacturing needs precise leveling. A microphone can typically be used to monitor the health and safety of such equipment.
100
Video cameras and vision systems are ?
smart sensors that have substantial processing power in the form of high-end processors, digital signal processors, FPGAs, and custom ASICs.
101
In modern vision systems, one of two types of sensing elements is used:
charge-coupled devices (CCD) or complementary metal-oxide (CMOS) devices.
102
In charge-coupled devices (CCD) the charge is moved from?
the sensor to the edge of the chip to be sampled sequentially via an analog-to-digital converter.
103
CCDs create:
high-resolution
and low-noise images.
Nevertheless, they consume considerable amount of power.
104
In the latter type, individual pixels contain transistors to?
sample the charge and allow each pixel to be read individually.
105
CMOS sensors are more common in the IoT market?
since they consume little power. However, they are mostly susceptible to noise.
106
The figure below demonstrates the various components of a typical CMOS camera and the various processing steps.
107
In general, CMOS sensors are integrated into ?
a silicon die that appears as a two-dimensional array of transistors arranged in rows and columns over a silicon substrate.
108
A sensor starts with a -------------that observes a scene. Afterward, an image --------------------------------------- is used to expose the captured image to a series of steps to -------------------------------------the image several times into a usable digital image.
lens
signal processor (ISP)
filter, normalize, and convert
109
The amount of data passing through?
a camera sensor at a conservative 60 frames per second and 1080p resolution is massive.
Hence, it is not recommended to stream this massive data to the cloud for processing.
110
As previously mentioned, most IoT applications require ?
the incorporation of several sensors to obtain higher-level inferences, thus making precise decisions.
111
sensor fusion.
The process of combining readings from different sensors is referred to as sensor fusion.
112
For instance,
a single temperature sensor has no clue of what causes a rapid temperature change since it has no contextual awareness.
113
Nevertheless, by combining these temperature readings with location detection and light intensity sensors, the IoT system could infer that a large number of people are congregating in a certain area while the sun is shining. In such a scenario, the IoT system may?
decide to increase air circulation in a smart building.
114
In general, there exist two major modes of sensor fusion:
centralized and de-centralized.
115
In the centralized mode,
the raw data is streamed and aggregated to a central service, e.g., cloud-based service, where fusion is performed.
116
In the de-centralized mode,
the data is correlated at the sensor or close, e.g., edge or fog nodes.
117
Autonomous vehicles stand as a prominent example of
de-centralized sensor fusion.
118
In fact, sensor fusion widely helps in?
constructing a hypothesis about the state of the environment the vehicle is in.
119
Above, an autonomous vehicle is depicted where ? Such inferred data leaves drivers safe since the autonomous vehicle becomes fully prepared for all road scenarios.
readings from five different sensors are fused together to track stationary and moving objects which is an important feature of autonomous driving technology.
120
In this scenario, data streams from several sensors are combined to?
estimate the position, velocity, trajectory, and class of objects, i.e., other vehicles and pedestrians.
121
The combined sensor readings in this application are as follows:
Ultrasonic sensors—to detect obstacles in the immediate vicinity
GPS—to calculate longitude, latitude, speed, and course
Speed and angle sensors—to measure speed and wheel rotation
LiDAR—to allow obstacles to be correctly identified
Camera—to detect, classify, and determine the distance from objects
122
The readings from these sensors are transmitted to?
a receiving equipment, such as an onboard processing unit.
123
One example of the fusion algorithms that can be implemented on a vehicle's processing unit is known a ?
“Kalman filter.”
124
Kalman filters work by ?
modeling the expected behavior of the sensors in response to movement, taking into account noise and distortions, and comparing this with what actually happens in order to iteratively refine the model until it accurately reflects the true behavior.
125
To this end, Kalman filters rely on both?
the probability theory and the collected measurements.
126
In general, a Kalman filter is an iterative process that uses two steps, referred to as?
predict and update.
127
Kalman filter
For simplicity, the mathematical formulas for the ‘predict’ and ‘update’ steps have been omitted.
128
For example
the control system of an autonomous vehicle may want to track a neighboring vehicle.
In the predict step, a Kalman filter forecasts the current position of this neighboring vehicle based on the previous position.
Moreover, it also computes the uncertainty in the predicted value.
129
An example of uncertainty may occur when?
the prediction model assumes that the neighboring vehicle is moving with a constant velocity due to zero acceleration. However, the neighboring vehicle will have acceleration, i.e., its speed will fluctuate from time to time.
130
the prediction model assumes that the neighboring vehicle is moving with ?
a constant velocity due to zero acceleration. However, the neighboring vehicle will have acceleration, i.e., its speed will fluctuate from time to time.
131
To refine its prediction, a Kalman filter exploits?
the various measurements from the attached sensors in the update step.
132
It calculates the difference between the predicted value and the measured values and then decides?
which value to keep through estimating the Kalman gain.
133
The output from the update step is ?
again fed back to the predict step and the process continues until the difference between the predicted value and the measured value approaches zero.
134
On one hand, Kalman filters can detect scenarios where?
data from one of the sensors is unreliable and should be ignored.
135
On the other hand, Kalman filters suffer from ?
a relatively slow reaction in the case of rapidly changing situations.
Moreover, they assume that both the system and measurement models’ equations are both linear which is not realistic in many real-life situations.
136
To overcome this shortcoming, other algorithms, such as ?
extended Kalman filter,
unscented Kalman filter,
and particle filter, must be adopted.
137
In general, power management is a significant component of any successful ---------------------------------------. Most IoT devices are empowered using --------------------- with limited energy ------------------------------. Accordingly, the expected lifetime of such devices is widely --------------------------------------.
IoT deployment
batteries
budget
limited
138
Accordingly, the expected lifetime of such devices is widely limited. To extend the lifetime, system designers must either?
adopt a method for collecting energy from the environment, i.e., energy harvesting,
or adopt an energy conservation method to wisely consume the available energy.
139
The most common energy storage facility for IoT devices is ?
either a battery or a supercapacitor.
140
To select the most suitable energy source, system designers have to consider several aspects, including:
(1) battery energy capacity;
(2) how the battery's energy will vary over time as it is discharged;
(3) is the IoT sensor in a thermally constrained environment that can affect the battery lifetime and reliability;
(4) is the battery deployed in an accessible place where it can be replaced; and finally (5) the maximum allowed weight of the battery.
141
These days, several types of batteries are on the market, including?
lithium-ion (Li-ion),
nickel-cadmium,
and nickel-hydride batteries.
142
Li-ion batteries become the standard form of?
power in mobile devices thanks to their high energy density.
143
In a Li-ion battery, :
lithium-ions physically move from the negative electrode to the positive.
During recharge, the ions move back to the negative region.
144
To predict the lifetime of a battery, Peukert's law is vastly accepted by many IoT system designers?
Peukert's law approximately estimates the change in capacity of batteries at different rates of discharge.
145
According to Peukert's law, the lifetime of a battery is?
equal to where t = c/ln is the capacity of the battery,
I represents the discharge current, and n is Peukert's exponent.
As the rate of discharge increases, the battery's available capacity decreases.
As the exponent n increases, we move further from a perfect battery to one that discharges faster as the current increases.
146
In fact, the discharging rate differs for different types of batteries. Figure 3.8 depicts a comparison between three well-known battery types, including Li-ion, lead acid, and Ni-Cd. As shown in the figure, Li-ion provides a nearly constant voltage over its life but has a steep drop toward the end of its storage capacity. Alternatively, lead acid and Ni-Cd have less voltage potential and a curvilinear degradation in power.
147
Supercapacitors are another form of energy storage in IoT applications. In a supercapacitor, energy is?
stored electrostatically on a plate, and does not involve any sort of chemical transfer of energy like Li-ion batteries.
148
One of the most important advantages of supercapacitors over batteries is?
their fast charging time, i.e., they can charge to their full potential in seconds.
Moreover, it is straightforward in supercapacitors to predict the remaining time power will be available.
149
In general, energy-harvesting devices efficiently
capture,
accumulate,
store,
condition,
and manage
this energy to ?
energize the deployed nodes.
150
energy-harvesting devices are increasingly attractive alternatives to ?
costly batteries which have a relatively limited lifetime.
151
Energy is usually everywhere in the various wireless sensor network (WSN) environments. This energy is available in the form of:
-mechanical energy—from sources such as vibration, mechanical stress, and strain.
-thermal energy—waste energy from furnaces, heaters, and friction sources.
-light energy—captured from sunlight or room light via photo sensors, photo diodes, or solar panels.
-electromagnetic energy—from inductors, coils, and transformers.
-natural energy—from the environment such as wind, water flow, ocean currents, and solar.
-human body—a combination of mechanical and thermal energy naturally generated from bio-organisms or through actions such as walking and sitting.
-other energy—e.g., from chemical and biological sources.
152
An energy harvesting system typically requires an
energy source
and two other key electronic components. ?
First, an energy conversion device, such as a piezoelectric element, translates the energy into electrical form.
Second, an energy harvesting module captures, stores, and manages the power for the device.
153
The selection between these energy sources highly depends on ?
the IoT application requirements and constraints.
154
However, IoT applications which leverage harvested energy to power their electronic components may have difficulties?
since the power availability may vary over time.
155
Furthermore, the available harvested power could be limited although --------the ---------------energy. The limited energy problem occurs due to the low ------------------------ of most energy conversion ---------.
not
surrounding
conversion efficiency
devices
156
Characteristics of various energy sources
Energy Source:
1-solar
Harvesting technology Conversion efficiency Amount of harvested energy
solar cells 15% 15 mW/cm2
2-finger motion
Harvesting technology Conversion efficiency Amount of harvested energy
piezoelectric 11% 2.1 mW
3-exhalation
Harvesting technology Conversion efficiency Amount of harvested energy
breath masks 40% 0.4 W
4-blood pressure
Harvesting technology Conversion efficiency Amount of harvested energy
micro-generator 40% 0.37 W
157
A prominent example of an energy transducer is?
the piezoelectric device.
Piezoelectric devices have been widely used as vibration sensors, but they can also be used to generate power.
158
The core idea is ?
to convert the mechanical strains to energy through motion, vibration, or even sound.
159
piezoelectric harvesters are suitable for very small systems with some form of energy collection and storage?
As shown in the table above, piezoelectric harvesters produce currents on the order of milli-watts.
160
Similarly, RF energy harvesting has been used for years in the form of?
RFID tags.
161
RFID exploits the near-field communication to send ?
RF power from the readers to RFID tags thanks to their proximity.
Alternatively, far-field applications involve harvesting energy from broadcast transmissions, e.g., televisions, cell signals, and radio.
162
The main limitation of such technology emerges due to:
- the rapid drop of power efficiency over distance.
-Moreover, active radiation is highly undesirable for health-related reasons. Below, we discuss some of the well-known methods for conserving the allocated energy.
163
During the design time, system designers must consider several power contributors while deciding upon the budget of the power source to be used. Examples of such contributors are:
the active sensor power (if any),
frequency of data collection,
wireless radio communication strength and power,
frequency of communication, microcontroller power as a function of core frequency,
passive component power,
energy loss from leakage or power supply inefficiency,
and power reserve for actuators and motors.
164
Taking these energy contributors into account during the design time leads to ?
the development of what is currently called as green IoT (G-IoT).
165
In other words, G-IoT is a --------------- of ------------------------------------- in which energy efficiency is a major key during the ------------------------------phases.
subset
IoT applications
design and development
166
In this context, G-IoT tackles the main energy consumers in the different components of IoT systems, e.g.,
sensing, communication, and data analytics.
167
For sensing, several strategies can be adopted to ?
avoid wasting energy.
168
For instance, IoT sensors can be adjusted to ?
work only whenever necessary, while spending most of their lifetime in a sleep mode.
169
To this end, efficient scheduling algorithms must be adopted to?
trigger the sensors in a timely manner.
170
Moreover, the communication protocols need to be flexible enough to deal with?
the network dynamicity which results from the frequent activation/deactivation of these IoT devices.
171
Another solution for the energy problem can be?
the local processing of the sensory readings at the sensors.
The core idea is to trade off processing for communication knowing that radio communication mostly consumes much more energy than data processing.
172
Accordingly, it is important to reduce the amount of data that must be reported to ?
the cloud services by means of fusion, compression, inference, or prediction algorithms.
173
As a new way of sensing the signal with a much lower number of linear measurements provided that?
the underlying signal is sparse, “compressive sensing” is also able to enhance energy efficiency.
174
In the field of green communication, cognitive radio (CR) stands as?
the most efficient method for improving the performance of wireless communication.
175
Why reducing the energy required for retransmissions. Specifically, CR systems enable various IoT devices from sharing the same communication channel?
because the most efficient method for improving the performance of wireless communication.
176
To this end, CR systems are aware of their -----------and can change their ---------of operation (e.g., ---------------------------------------------------------------------------) via software and hardware manipulation. Accordingly, they can improve ----------------------------------- efficiency and minimize the problem of spectrum -----------------------.
environment
modes
operating frequency, modulation scheme, waveform, and transmitting power
spectrum-usage
over-crowdedness
177
Sensors typically generate a massive amount of?
data in most IoT applications.
178
the cloud services cannot process big data generated from a myriad of IoT applications?
Transporting this big data to the cloud for processing could be a difficult task.
Specifically, sending all these data to the cloud requires excessively high network bandwidth.
Moreover, the cloud services are still limited in their capacity
179
the cloud services cannot process big data generated from a myriad of IoT applications . To overcome these challenges:
fog computing comes into play.
180
The term fog computing was coined by ?
Cisco in 2012.
181
Through this technology, several services can be efficiently provided to?
the IoT users, such as data processing and storage.
182
The core idea behind fog computing is?
to provide data processing and storage capabilities close to the sensors instead of sending them to the cloud.
183
In this manner, fog computing:
-improves efficiency and performance,
-and reduces the amount of data transferred to the cloud for processing, analysis, and storage.
184
In fact, processing and storing data on edge processors, close to the sensors, highly reduces?
network traffic and latency.
185
The figure below delineates fog computing as an intermediary layer between?
the cloud and end devices.
186
Such fog devices bring processing, storage, and networking services closer to ?
the end devices themselves.
187
In fact, any device with computing, storage, and network connectivity can be?
a fog node,
e.g.,
industrial controllers, switches, routers, embedded servers, and video surveillance cameras.
188
In general, the primary characteristics of fog computing are summarized as follows:
-location awareness. Fog computing supports location awareness in which fog nodes can be deployed in different locations.
-geographical distribution. In contrast to the centralized cloud, the services and applications provided by the fog are distributed and can be deployed anywhere.
-support for mobility. One of the important aspects of fog applications is the ability to connect directly to mobile devices and therefore enable mobility methods.
-real-time interactions. Fog computing applications provide real-time interactions between fog nodes rather than the batch processing employed in the cloud.
-interoperability. Fog nodes are designed by different manufacturers, thus coming in different forms. They can interoperate and work with different domains and across different service providers.
189
The fog computing hierarchy is ?
one in which a network of fog nodes and the cloud services are connected and they are divided into logical or physical tiers .
190
The fog computing hierarchy is composed primarily of three main tiers:
such that each tier assigns special duties to its corresponding devices.
In other words, each tier holds special responsibilities different from the others.
191
Generally, the lower tiers, which are the ones closest to the IoT devices, are ?
the ones responsible for the latency-sensitive data,
and as you go up the hierarchy, the responsibility is directed more toward bigger, memory-consuming, and latency-insensitive data.
192
The three tiers are as follows:
a-monitoring and control. This is the closest tier to the IoT devices. Its main function is to react upon generated data from the sensors and change the state of the device based on that received data. It has very low storage capability and geographically covers a very local area, but it offers the fastest servicing among all tiers. Several terminal nodes read different physical parameters and act upon them in real time. They can sense problems and command actuators to respond effectively to these problems. Moreover, each terminal node can share its absolute geographical location so that they can support location-based services.
b-operational support. It is composed of nodes that are usually called aggregation nodes. Their primary function is to aggregate the streamed data from the tier below them for data analytics, filtering, compression, transformation, and responsive actions in near real time. Each aggregation node collects its data from one or more lower-level fog nodes, depending on the application. They have more storage power than the lower-level fogs and cover a larger area, but their latency is therefore higher. Devices that comprise this tier are usually smart intermediate devices with some storage, computation, routing, and packet forwarding abilities, such as routers and switches.
c-business support. It is the tier closest to the cloud service or even the cloud core itself. Its main function is to turn aggregated data into useful knowledge to provide enterprise level services for businesses, as it has got the highest storage and analytical abilities. This tier is the centralized tier to which all the fog nodes of all the tiers below are connected and to which they transmit their data. It is composed of many high-end servers and data centers capable of fulfilling its duties. Nevertheless, its designated nature implicates the highest latency in the whole hierarchy. For that reason, data is transmitted there only when necessary which characterizes the efficiency of fog computing in terms of resource allocation and packet forwarding. The following table summarizes the different features of the three tiers.
193
Fog computing tiers
see the table
194
The hierarchy is designed to be ?
scalable according to the business needs, in terms of storage, network connectivity, and analytical services.
195
A given business model does not necessarily need to implement a three-tier architecture. In other words?
the three main functions fulfilled by the three main tiers do not necessarily need to be implemented in separate devices.
Some use cases may require all three of them to be done in the cloud.
Other use cases may require only the monitoring and control tier to be implemented in the fog and the other two tiers in the cloud, and so on.
196
The architecture is designed to be?
totally flexible in terms of which functions should be performed by, or deployed on, which devices.
197
The following figure shows four different scenarios that require four different deployments of the hierarchy.
198
The three tiers of the stack must be deployed on fog nodes?
The first scenario is entirely cloud-independent.
Such a use case is assumed to be extremely latency-sensitive.
199
In this context, incorporating the cloud is without a doubt irrational. Examples for such a scenario are:
health care systems, ATM banking systems, and armed forces combat systems.
200
Such applications need to be designed as deadline-critical systems.
Apparently, in all three examples we are dealing with humans’ lives.
Another possible reason for using this scenario is the unavailability of the cloud within a particular geographical location, which would make it impossible to involve the cloud.
201
The second and third scenarios are hybrids of both?
fog nodes and the cloud.
202
The second and third scenarios are more common in ?
the IoT domain.
203
The second scenario exploits ---------------- in the first and second tier, while the third scenario uses them only in the---------------. This implies that the second scenario is suitable for the more ------------------- use case as compared to the third scenario.
fog nodes
first tier
latency-sensitive
204
The second scenario supports several use cases, including :
commercial building management, commercial solar panel monitoring,
and retail,
205
whereas the third scenario supports use cases such as?
commercial uninterruptible power supply (UPS) device monitoring,
mobile network acceleration,
and content delivery networks (CDNs) for Internet acceleration.
206
The fourth scenario, on the contrary, is ?
totally fog-independent,
where it might not be feasible or economical to use the fog in such use cases.
207
Examples of such cases include ?
agriculture,
connected cars,
and remote weather stations.
208
It is worth mentioning that while those three tiers form the basis of the fog computing architecture, businesses can?
add more layers to the hierarchy if they sense the need for extra intelligence in their system.
209
businesses can add more layers to the hierarchy if they sense the need for extra intelligence in their system?
-The more layers they add, the deeper the data mining and analysis and, thus, more intelligence.
210
Moreover, fog nodes are built with an intelligence that sometimes, if a fog node is busy with processing data, could?
redirect the newly received data to other fog nodes.
211
The other fog nodes can be?
in the same tier or even in a higher tier.
212
Each fog node is equipped with this intelligence and insight to determine?
whether to redirect data to another fog node or to the cloud in the higher tiers.
213
In the basic mode, the terminal nodes are deployed on?
the sensors embedded in IoT devices, where they can sense and share their geographical location.
214
The localized terminal nodes are ?
-grouped together to form a location-based logical virtual cluster (VC) that is seen as one unit by the tier above.
-Each VC gets connected to its geographically nearest fog instance (FI), and through edge gateways, data transmission is possible between both tiers.
215
The second tier in the hierarchy is ?
composed of fog-aggregation nodes that can be deployed on networking devices that bear some computing, analytical, and storage feature, e.g., routers and switches.
216
Each device embodying a fog node is referred to as an ?
FI. Each FI can analyze, process, and store some amount of data, depending on its designated function within the hierarchy.
217
Through fog gateways, FIs become able to transfer?
data to the next and final tier, the cloud tier.
218
Each fog gateway can be connected to one or more cloud gateway which are?
additionally connected to the mesh of cloud data centers (DCs).
219
These DCs are designed to handle?
huge data in terms of processing power and permanent storage with their superb computational, analytical, and storage capabilities.
220
Alternatively, the SDN-based architecture is designed to ensure?
the system flexibility .
221
This is a kind of networking whose basic concept is to?
adapt a control plane within the network that separates control functions from data-forwarding functions.
222
This is a kind of networking whose basic concept is to adapt a control plane within the network that separates control functions from data-forwarding functions. This separation helps:
both facilitate packet forwarding within the network (i.e., making it more efficient, flexible, and scalable) and provide better quality of service (QoS) management for enterprises.
223
An SDN is typically composed of a data plane, in which switches forward the data packets, and the control plane which carries?
the network controller and management operator to control how the devices in the data plane operate.
224
Through an application programming interface (API) that the controller provides, the control panel is?
programmed to function in a desired way affecting the whole network.
225
The SDN-based architecture is depicted below. The core idea is?
to exploit the massive number of cellular base stations (BSs) deployed in the mobile network on a global level.
226
These BSs could be utilized in?
an IoT network to form the means for implementing a more flexible fog computing architecture.
227
Each BS could be directly connected to?
a fog node that aggregates data from local devices, processes some of it, and sends the rest through an SDN-cellular core to the cloud.
228
Each BS group of a particular location could be connected to ?
an edge device in the network core.
229
This edge device can also be connected to?
a fog node to aggregate data from the localized base stations that are connect to it.
230
The SDN-based architecture adopts ?
OpenFlow switches in the network, and an OpenFlow controller operates all the base stations and switches via the OpenFlow Protocol—an SDN protocol developed by the open networking foundation (ONF) in 2011.
231
The OpenFlow controller manages ?
the forwarding plane of the base stations and the switches while monitoring the network traffic.
232
The controller can be reprogrammed through an API by businesses to ?
add or remove features to or from the network, thus providing flexibility of functioning.
233
Before delving into the various platforms, it is necessary to define some factors for properly selecting the most suitable hardware platform for the targeted IoT application. Such factors include:
the processor speed,
memory requirement,
networking facilities,
and power consumption.
234
The processor speed indicates?
how fast it can process the captured sensor readings.
235
Naturally, a faster processor speed means?
that it can execute instructions more quickly.
236
The clock speed is still the ------------------------------------for raw computing power, but it is ---------------the only one. A comparison among various -------------------------------can be done using the millions of ----------------per second (MIPS) metric.
simplest proxy
not
platforms
instructions
237
Some processors may?
lack hardware support for floating-point calculations.
238
If the code involves a lot of complicated mathematics, a by-the-numbers slower processor with hardware floating-point support could be?
faster than a slightly higher performance processor without it.
239
Generally, the processor speed is used as ?
one of a number of factors when comparing similar systems.
240
Microcontrollers tend to be clocked at speeds in the tens of MHz, whereas system on chip (SoC) runs
at hundreds of MHz or possibly low GHz.
241
If the IoT application does not require heavyweight processing—for example, if it needs only networking and basic sensing—then ?
some sort of microcontroller will be fast enough.
242
If the device will be crunching lots of data—for example?
processing video in real time—then SoC platforms are recommended.
243
Aside from the processor speed, the memory footprint is?
another important factor for selecting the most suitable IoT platform.
244
In this context, random access memory (RAM) provides ?
the working memory for the system.
245
If an IoT platform has more RAM, the device may be able to perform more?
processing or have more flexibility over your choice of coding algorithm.
246
It is difficult to give exact guidelines to the amount of RAM that may be required by ?
the IoT platform, as it will vary from one application to another.
However, microcontrollers with less than 1 KB of RAM are unlikely to be of interest,
and if you want to run standard encryption protocols, system designers will need at least 4 KB, preferably more.
247
For SoC boards, particularly if the system designers plan to run Linux as the operating system, at least --------------------- is recommended.
256 MB
248
The IoT platforms are also expected to?
be connected to the Internet.
249
Wired Ethernet is?
often the simplest technique—generally plug and play—and cheapest, but it requires a physical cable.
250
Wireless solutions obviously avoid that requirement but?
introduce a more complicated configuration.
251
Wi-Fi is the most widely deployed solution to provide an existing infrastructure for connections, but ?
it can be more expensive and less optimized for power consumption than some of its competitors.
252
Other short-range wireless options can offer better power-consumption profiles or?
costs than Wi-Fi, but usually with the trade-off of lower bandwidth.
253
ZigBee is?
one such alternative technology, aimed particularly at sensor networks and scenarios such as home automation.
254
The recent Bluetooth LE protocol (also known as Bluetooth 4.0) has?
a very low power-consumption profile similar to ZigBee’s and could see more rapid adoption due to its insertion into standard Bluetooth chips included in phones and laptops.
255
There is an existing Bluetooth standard as another possible choice?
For remote or outdoor deployment,
little beats use the mobile phone networks.
For higher data rates,
data connections, such as 3G, and LTE communication networks can be used.
256
In terms of energy consumption, faster processors are?
typically more power hungry than slower ones.
257
For the IoT devices which might be portable or rely on an unconventional power supply (e.g., batteries, solar power), power consumption could be?
an issue.
258
For the IoT devices which might be portable or rely on an unconventional power supply (e.g., batteries, solar power), power consumption could be an issue. Even with access to main electricity, it may be necessary to consider ?
because lower consumption may be a desirable feature. Nevertheless, processors might have a minimal power-consumption sleep mode.
259
power-consumption sleep mode can enable the system designers to use?
a faster processor to quickly execute tasks and then return to the low-power sleep mode.
260
Accordingly, the utilization of powerful processors may not be?
a disadvantage even in a low-power embedded device.
261
In general, IoT hardware can be classified into two broad categories:
Now, we discuss some of the most popular platforms and boards that can be used by IoT developers for initial prototyping, creating smart objects, and developing projects and products.
(1) wearable devices and gadgets and
(2) embedded systems and boards, as depicted below.
In the former category, many preassembled standard hardware applications, ranging from smart shoes to glasses, are available.
262
The scope of IoT development in this category is limited to ---------------------. On the other hand, both the hardware and software aspects are open for system -------------------------under the ---------------------systems and boards category.
software
designers
embedded
263
ESP8266 is?
a popular Wi-Fi solution for connecting an edge to the Internet. It has two primary versions with different capabilities.
264
The first version is?
the generic ESP8266 module which can be easily interfaced with various microcontrollers.
It has 1 MB of flash memory, works on the 802.11 b/g/n protocol, and supports Wi-Fi Direct (P2P) and soft-access point.
265
It comes with an integrated transmission-control protocol/Internet protocol stack and a self-calibrated RF antenna which allows it to?
operate under almost all conditions.
266
The second module is?
=the SparkFun Blynk Board which is an ESP8266-based board with nine general-purpose input/output (GPIO) pins supporting serial-peripheral interface (SPI) and inter-IC (I2C) communication protocols.
=It has an onboard lithium-polymer (Li-Po) battery connector and charging port.
=It also comes with onboard Future Technology Devices International for reprogramming, red-green-blue light-emitting diode (RGB LED), analog-to-digital converter (ADC), and temperature and humidity sensor.
= It can automatically connect to Blynk Cloud and can be controlled with the Blynk app which is available for both the iOS and Android operating systems.
267
The second module Despite being relatively costly, this module has been heavily used for?
home automation applications.
268
Arduino boards are?
another popular type of IoT platform capable of performing tasks from blinking an LED to publishing material online to handling heavy networking tasks.
269
In general, Arduino has a broad range of boards, from simple-----------------------------boards to products for ----------------------, three-dimensional (3D) printing, and much more. For supporting IoT applications, the Arduino ----is supported with onboard Wi-Fi (IEEE 802.11 b/g/n) and Ethernet (IEEE 802.3 10/100Mb/s).
8-b microcontroller
wearables
Yun
270
It has an ATmega32u4 with a clock speed of 16 Mhz and an Atheros AR9331 (MIPS @ 400 MHz) which backs a Linux distribution called OpenWrt-Yun. It also includes a Li-Po charging circuit and comes with Cryptochip for secure communication. Thanks to these features, these modules have been used in several IoT applications, such as?
computationally complex applications and algorithms,
high-speed communications,
telemetry hubs that gather data wirelessly from sensor nodes,
and many other IoT-based applications.
271
Raspberry Pi mainly has two different generations.
The first generation,
referred to as “Raspberry Pi 2,” comes with a single-board computer that has a quad-core, a graphics processing unit, 1 GB of RAM, 40 GPIO pins, 4 USB 2.0 ports, 1 Ethernet port, 1 HDMI connector, and 1 micro-SD card slot.
Alternatively, the second generation,
called Raspberry Pi 3, is ten times faster than Raspberry Pi 2. It is powered with a 1.2-GHz 64-b quad-core ARM CPU and has integrated 802.11n wireless LAN and Bluetooth 4.1. With the addition of onboard wireless LAN and Bluetooth capability, it will be even more useful for developing IoT applications.
272
Raspberry Pi can run many operating systems, including:
Raspbian Linux,
Ubuntu Mate,
and Windows 10 IoT Core.
273
It also provides full support for such programming languages as ?
C/C++, Python, and JavaScript.
274
Raspberry Pi boards for IoT projects are ?
an excellent choice.
However, additional hardware is required for interfacing with analog inputs such as potentiometers, photocells, and joysticks, as ADC is not available onboard.
Moreover, the board consumes slightly more power compared to earlier versions and gets a bit warmer when used in the overclocking mode for a longer time.
275
Developers using Raspberry Pi have come up with many intriguing IoT projects, such as?
Rebroadcast Internet Radio and an Internet weather station.
276
The manufacturer comes with? Once cloudBit is connected to Wi-Fi, it starts sending data from other littleBits modules to the cloud without a need for programming. Moreover, it is supported by many APIs. The ease with which every module can be magnetically attached to others makes it very popular among IoT developers.
almost 60 interchangeable modules that are attached to each other magnetically in billions of possible combinations.
277
The manufacturer comes with?
almost 60 interchangeable modules that are attached to each other magnetically in billions of possible combinations.
278
The cloudBit is one of?
the 60 bits or modules. It comes with a Linux-based system on a Freescale ARM processor with 64 MB of RAM. It makes use of an 802.11 b/g/n USB adapter for networking.
279
Once cloudBit is connected to Wi-Fi, it starts?
sending data from other littleBits modules to the cloud without a need for programming.
Moreover, it is supported by many APIs.
280
The ease with which every module can be magnetically attached to others makes it ?
very popular among IoT developers.
281
The main components of a typical IoT system architecture are:
sensors, data communication, internet and routing protocols, cloud and fog computing, IoT security.
282
Developing an IoT application requires a very multidisciplinary know-how ranging from sensor technologies to machine learning?
Because the main components of a typical IoT system architecture are: sensors, data communication, internet and routing protocols, cloud and fog computing, IoT security.
283
Sensors denote any device which can?
generate data by sensing the world.
284
Depending on the application, the generated amount can be?
huge or small.
285
novel energy supply methods such as energy harvesting should be adopted.
To be affordable, sensors need to be low-size and low-costs. Moreover, they need power to operate.
286
The data communication refers to ?
the moving of generated data, usually from sensor nodes, to dislocated cloud services and data centers.
287
There is an heterogeneous technology base to?
perform data communication.
288
The adoption of the right one yields?
a better data quality and performance of the overall IoT system.
289
The adoption of the right one yields?
a better data quality and performance of the overall IoT system.
289
The adoption of the right one yields?
a better data quality and performance of the overall IoT system.
290
Internet and routing protocols target the bridging of data from sensors to the Internet through?
gateway routers and devices supporting IP-based protocols.
291
IoT data need efficient low-latency protocols that can be?
steered and secured in and out of the could.
292
Cloud and fog computing deal with ?
processing the collected data.
293
Here a tradeoff between two approaches is to be found based on?
the application scenario and the available computing resources.
294
On the one hand, all data can be?
moved to the cloud for further processing.
On the other hand, the data can be processed at the edge (edge computing)
or extending cloud services downward into an edge router (fog computing).
295
In many applications a local processing can be?
meaningful to easily and fast detect some patterns in the data.
296
In many applications a local processing can be?
meaningful to easily and fast detect some patterns in the data.
297
Security is ?
the last important component of an IoT system architecture.
298
Technologies such as blockchain and software-defined perimeters can enhance?
the robustness of the IoT application against attacks.
