1
Multicriteria Selection of Transmission Parameters in the IoT
Sinda BOUSSEN1, Mohamed-Aymen CHALOUF2 and Francine KRIEF3
1 Mediatron, University of Carthage, Tunis, Tunisia
2 IRISA, University of Rennes 1, Lannion, France
3 LaBRI, Bordeaux INP, Talence, France
1.1. Introduction
Cognitive radio networks (CRN) and the Internet of Things (IoT) are concepts that are taking on more and more importance in modern communication systems. The IoT extends connectivity to the real world (Atzori et al. 2010; Perera et al. 2014, 2015; Xu et al. 2014; Al-Fuqaha et al. 2015), by allowing an object (a sensor, smartphone, car, etc.) to interact with the existing internet infrastructure, to communicate with other objects and to collect and exchange data. An object can have several interfaces and so detect different access networks. When several access networks are available, it will be necessary to make a decision about selecting the access network best adapted to the current situation. However, with the increase in the number of connected objects, the spectrum is becoming a precious resource, one threatened with scarcity. Faced with this problem, we can imagine objects with cognitive capacities (Vlacheas et al. 2013). This will make it possible to optimize use of unoccupied radio frequencies while still minimizing interferences with priority users. In an intelligent radio environment (Haykin 2005; Akyildiz et al. 2006), we distinguish two types of users: priority users, called primary users, who have an exclusive right over some of the spectrum, and secondary users, also called cognitive users, who have “opportunistic” access to the spectrum. The main functions of intelligent radio are spectrum detection, decision-making, spectrum sharing and spectrum mobility. There is a great deal of research focusing on IoT architectures based on intelligent radio (Shah et al. 2013; Vlacheas et al. 2013; Aijaz and Aghvami 2015; Khan et al. 2016), which has demonstrated the need to integrate intelligent radio into the IoT (Wu et al. 2014a).
In this chapter, we tackle the question of making a decision for effective access to an access network or radio channel in the IoT. To communicate, an object that has several interfaces and/or cognitive capacities can detect numerous access networks or radio channels. In this case, it is necessary to choose the access network or radio channel best adapted to the communication in question. Among those aspects that might be considered in this decision, we might find Quality of Service (QoS) constraints on the IoT application and the object’s own energy constraints (Zhu et al. 2015; Shaikh et al. 2017). Thus, the decision-making module should be multicriteria and consider the dynamic context of the radio environment, the needs of the application in terms of QoS and the object’s own energy constraints.
The remainder of the chapter is organized as follows. Sections 1.2 and 1.3 give an overview of recent work on managing vertical handover and spectrum handoff, respectively, in an IoT context. In section 1.4, we suggest a multicriteria decision-making module that makes it possible to select and adapt radio transmission perimeters in order to satisfy the QoS requirements of the transported data equally as well as the energy constraints of the communicating objects. The suggested module will be used equally as much where heterogenous networks co-exist as in the context of intelligent radio. The choice of access network (presence of several heterogenous networks) or radio channels (use of intelligent radio) depends on the number of types of contextual information such as the networks state, user preferences, the application constraints and the characteristics of the object. To evaluate the suggested decision-making module, we consider the domain of Vehicular Ad-hoc NETworks (VANETs) and examine two possible situations: with and without intelligent radio.
1.2. Changing access network in the IoT
Managing mobility, QoS and energy in heterogenous networks is a major challenge for modern communications systems. So, a number of offers have come to light in recent years to enable a user to stay connected as much as possible, in any place.
In cases where heterogenous networks co-exist, an object equipped with several interfaces should be able to choose the access network best adapted to its needs, from among those available. Vertical handover (VHD) decision-making algorithms were designed to this end. The main purpose of these algorithms is to ensure the required QoS while still guaranteeing a transparent mobility between the different access technologies. Recently, several solutions have been suggested to overcome these challenges in an IoT context. These solutions (VHD algorithms) can be classed into five categories (Kassar et al. 2008; Zekri et al. 2012; Bhute et al. 2014) depending on the approach considered: classical, artificial intelligence (IA), cost-dependence, multiple attribute (Xiao and Li 2018) and decision-making, depending on the context.
In addition to the constraint linked to QoS and to mobility, the objects have an energy constraint. It is therefore important to consider energy consumption when selecting the most appropriate access network. Thus, several VHD solutions focus on energy efficient networks (Tuysuz and Trestian 2017). Most of these solutions used IEEE 802.21 MIH (Iqbal et al. 2019) and ANDSF (Access Network Discovery and Selection Function) protocols to collect information. Moreover, existing VHD systems make it possible to save energy by minimizing the scanning/detection time needed to discover the wireless network or to select the access point most economical with energy (Xenakis et al. 2011).
1.3. Spectrum handoff in the IoT
To avoid the scarcity of frequencies introduced above, the IoT can use the concept of intelligent radio. Many researchers have studied problems linked to spectrum management and channel allocation (Wu et al. 2014b; Koushik et al. 2018; Tarek et al. 2020). In Kumar et al. (2016), the authors studied spectrum handoff schemas in a CRN context and identified three types of transfer. First of all, the reactive approach, where the spectrum handoff and reconfiguration of the radio frequency happen after the primary user has been detected. Then, there is the proactive approach, where the spectrum handoff and frequency reconfiguration take place before a primary user has occupied the channel. Finally, the hybrid approach combines the two previous approaches. Recently, intelligent radio has been used in the IoT, especially in the two vehicular ad-hoc networks. This use has made it possible to improve the network’s spectrum efficiency and the experience of itinerant users by optimizing vehicle communication capacities and the QoS of applications such as road safety and traveler entertainment (Singh et al. 2014; Kumar et al. 2017).
To ensure continuity of service, we have opted for a proactive approach using a prediction module tasked with calculating the probability of future channel availability as well as the average time it is available.
1.4. Multicriteria decision-making module for an effective spectrum handoff in the IoT
In the IoT, the multicriteria solution seems most appropriate for making decisions about VHD, and by considering a great deal of contextual information to ensure effective selection of an access network/radio channel. To overcome the complexity of implementation, we opt for a cost function solution that calculates different candidates’ scores (access network or radio channel) and selects the one with the highest score. In the approach retained, we focus both on the QoS and on energy consumption, with the weight attributed to these parameters depending on the general communication context (network, application, object and user). Contextual information from the network is collected with the help of the surveillance module (context of multiple access networks) and of the intelligent radio-detection module (CRN context).
1.4.1. General architecture
In this section, we detail the general architecture we propose (Figure 1.1). In the approach we have retained, based on costs, the decision-making mechanism makes it possible to select the access network best adapted in the case of an object with several network interfaces and, on the other hand, to adapt the transmission parameters (channel, frequency, modulation, etc.) in the case of an intelligent radio network context. This selection will be based on the information available about the application, the user, the current radio conditions and predictions about how they will evolve. Information on the radio conditions is provided by the radio channel detection module (Akyildiz et al. 2006) in the CRN context and by the network monitoring module in the context of Multiple Radio Access Networks (M-RAN). The multicriteria decision-making module and the other general architecture modules are implemented within the object, which may be a car in the case of vehicular networks.
Figure 1.1. Proposed architecture for a context aware IoT device/object. For a color version of this figure, see www.iste.co.uk/chalouf/intelligent.zip
1.4.1.1. Detection for an intelligent radio module
In the context of CRNs, the detection module provides information on the accessible radio channels, their quality and their rates of occupation. A candidate radio channel (noted in the CR channel suite) is characterized by a central (fixed) frequency, a passband width (fixed), modulations and schemas for possible codings (fixed), a probability of availability (predicted), an average availability time (predicted), an energy cost (estimated), a packet loss rate (estimated), a passband (fixed), an average delay (estimated) and an average jitter (estimated).
1.4.1.2. Prediction module
The radio environment is not stable and the radio parameters may change for many reasons such as mobility and the arrival or departure of other objects. Thus, an object may need to change access network or transmission parameter values. To avoid delays and disruptions to transmission, unusable spectrum handoffs, we use a prediction module. This module will predict future variations of some parameters to anticipate decisions on spectrum handoff and carry this out at an opportune time. In the CRN context, a period could be equal to the time interval allocated to an object to carry out spectrum detection operations, signaling and the transfer of data along a given channel. Thus, when a degradation of the QoS is perceived for the current period (t), the decision to change channel is taken for the future period (t + 1).
At each period t, the prediction module calculates and predicts some transmission parameters for the period t + 1. If the transmission conditions are satisfactory for the period t + 1, then it is not necessary to change channels. Otherwise, the object considered should modify the transmission parameters to guarantee the required QoS.
The process of prediction may require significant resources (CPU, memory and energy). Consequently, implementation of the prediction module in the object may be very restrictive, indeed impossible. Thus, we suggest migrating the prediction process to the Fog radio access network. This makes it possible to improve the object’s capacity, to economize on the battery and to provide the user with a better experience.
The parameters that will be predicted at the period t + 1 are channel availability and the average channel availability time.
1.4.1.2.1. Probability of channel availability at period t + 1
In a CRN context, according to the analytical model defined in Song and Xie (2012), the predicted probability that a candidate channel i is inactive at time t can be expressed as follows (equation [1.1]):
where 
represents the arrival time of the kth packet and designates the length of the kth PU packet (data from one primary user) on the channel i. Ni(t) represents the state of the channel, which is a random binary variable with values 0 and 1 representing, respectively, the inactive and the occupied states.
The probability that the channel i is in active at an instant t can be obtained by equation [1.2]:
where 
represents the interarrival of the PU packets. Supposing that the PU packets arrive following a Poisson process (Wang and Wang 2008), distributed exponentially with the average packet arrival rate per second λi, the length of the PU packet following fLi (l).
Based on the prediction above, the candidate channel i is considered inactive for the following time interval when Pr (Ni (t) = 0) ≥ τH, where τH is the probability threshold for a channel to be considered inactive at the end of the current frame.
1.4.1.2.2. Average channel availability time
In an intelligent radio context, we used the model suggested by Song and Xie (2012) to predict the average channel availability time at instant t. The OFF period represents the period the channel is unavailable. For the ith channel, the cumulative distribution function (CDF) over the duration of the OFF period is given by equation [1.3]:
where toff represents the duration of the OFF period.
In order to support at least one SU packet (data from a secondary user), the probability that the inactive duration of the ith channel is longer than a transmission time of the frame η should be greater than or equal to θ (see equation [1.4]):
where θ is the probability threshold for a channel to be considered inactive for the next packet transmission.
1.4.1.3. Object characteristics
An object is characterized by the emission power PTx and state of its battery. The battery state or its remaining life (expressed in hours) can be calculated (Grace et al. 2009) on the basis of the battery’s capacity, its tension, its emission power and the constant K, which represents the power required for signal treatment.
1.4.1.4. Network monitoring module
In the M-RAN context, the network monitoring module provides information on the available access networks. These available access networks have different characteristics in terms of bandwidth, delay, loss rate, channel occupation rate and energy expenditure. The information collected (Carneiro et al. 2009) is analyzed to take a specific decision based on the services required.
1.4.1.5. Database
The input metrics needed by the suggested multicriteria decision-making module include the characteristics of the access network or radio channel, the characteristics of the object, application requirements, users’ QoS preferences and the energy constraints and user sensitivity to some bands that could interfere with medical equipment in an e-health context, for example. All this information is stored in the databases. We distinguish types of input parameters: the measured, predicted and static parameters. The detection module for the intelligent radio and the monitoring module in the M-RAN context provide measured parameters. The prediction module will be responsible for predicting channel availability and the average duration for which they are available. The static parameters include fixed characteristics of the network or radio channel, user preferences and application requirements. The application needs we retain are linked to the QoS required by the application, such as the minimal bandwidth, the maximum delay allowed, the maximum jitter tolerated m and the maximum packet loss rate tolerated.
1.4.1.6. Multicriteria decision-making module
Depending on the needs of the application (data type, transmission urgency), the spectrum availability and the remaining battery life, the multicriteria decision-making module will select the most appropriate access network or radio channel to ensure transmission. This module considers both the energy constraints and the QoS and will be detailed in section 1.4.2.
To ensure decision stability, which is an important criteria (Wang et al. 2014), and to avoid inefficient or unhelpful transfer decisions in both the contexts considered (M-RAN or CRN), we suggest using a waiting time (timer) between the selection and execution stages. At the end of this time, the device checks that the following conditions are still valid: (1) the network/candidate channel is still detected by the object, and (2) this candidate’s score (network/channel) is still higher than that of the current network/channel. If these conditions are verified, then the transfer (network or channel) is run, otherwise it is canceled and the selection process restarts from the beginning.
As this wait can be an obstacle to continuity of service, a compromise needs to be found. So, we suggest adjusting this waiting time depending on the context of the decision.
1.4.1.7. Module for running vertical handover/spectrum handoff
The running stage refers to establishing the connection with the new access network (M-RAN context) or changing channel/transmission parameters (CRN context). In the M-RAN context, three cases are possible so long as the old link is released before (Hard Handover), during (Seamless Handover) or after (Soft Handover) the new link is established.
1.4.2. Decision-making flowchart
The decision-making module, detailed in section 1.4.1, makes it possible to select the most appropriate access network (M-RAN) or the radio channel (CRN). This selection depends on many contextual parameters, which include characteristics of the radio environment, application needs, object capacities, etc. In our case, selecting an access network or CR channel means selecting all the parameters that characterize transmission. In fact, an access network can be considered as one candidate channel among others. This is a key element for guaranteeing that our suggestion functions in all possible cases (as Figure 1.2 shows). The n candidate networks/channels detected are written C1, C2, C3, …, Cn. On each channel (Ci), it is possible to transmit with the help of m modulations M1, Mj, Mm. Each Ci channel (Figure 1.2) is characterized by the probability of availability Ai, the occupation rate Oi, the bandwidth, the delay, the loss and the energy cost. Each channel is also characterized by one or more transmission models (MODi,m, where i designates the channel and m designates the modulation and coding).
Figure 1.2. General context of decision-making
The suggested multicriteria decision-making module is based on the concept of utility to reduce the complexity of such a decision. The utility function used to calculate the scores is U (x) =1− e–αx with x the parameter of the decision vector and α its corresponding weight (α > 0). For x ≥ 0, U(x) ∈ [0, 1].
The suggested selection algorithm is formed of the following stages. The first stage consists of verifying the availability of the different channels detected to eliminate poor candidates that have an availability probability below a defined threshold or whose availability time is less than a given duration. This preselection makes it possible to avoid unhelpful score calculation. At the second stage, the channels selected during the first stage are classed according to their energy scores. At the third stage, the decision-making module calculates the QoS score for each selected channel. After having calculated scores linked to QoS and to energy, the total score is calculated for each channel selected depending on the weight given to each aspect considered: QoS and energy. This is stage 4. The attribution of weight to criteria should reflect the importance of each decision-making criteria depending on the overall context (application, battery, etc.). Finally, the selected channel will be the candidate with the highest total score.
1.4.2.1. Stage 1: selecting available access networks/CR channels
In the M-RAN context, this processing stage is very simple. In fact, each access network detected is available. However, in a CRN context, this stage requires the following processing. First of all, it involves checking the availability of the current channel k and determining the probability Pr(Ck (t) = 0) that the channel k will be inactive at instant t. The secondary user should switch to a new channel when the availability of the current channel is lower than the probability threshold (equation [1.5]):
where Ck (t) represents the current channel k, which is a binary random variable with the values 0 and 1 representing the inactive and the occupied states, respectively.
τL is the probability threshold below which a channel is considered occupied and the secondary user should be allocated a spectrum handoff (Song and Xie 2012), that is, the current channel is no longer considered inactive at the end of transmission from the frame.
Then, for each channel i detected, it remains to determine the probability of inactivity and the average availability of the channel for the next period.
The candidate channel i is considered inactive for the next period when Pr (Ci(t) = 0) ≥ τH, where τH is the probability threshold for a channel to be considered inactive at the end of the current frame (Song and Xie 2012). Thus, the preselected channels are those that have a probability 
and an average duration of availability higher than a threshold 
tiOFF represents the duration of the OFF period of the channel i.
To support at least one SU frame, the probability that the activity duration of the ith channel will be longer than one transmission time of the frame η should be higher than or equal to θ. θ is the probability threshold for a channel to be considered inactive for the next frame transmission (Song and Xie 2012).
Thresholds τL and τH are not fixed arbitrarily. The choice of these thresholds can have an impact on the number of channels filtered during this stage. For this, we will choose both these thresholds depending on the radio context, again, such as the noise that characterizes this environment or the inactivity of primary users. For example, when there are very few channels considered available, we can lower the value of τL = τH. On the contrary, if these channels are very numerous, the value of τL = τH will be increased further. This will make it possible not only to select the channels with a high probability of availability, but also to reduce the number of candidates for which the decision-making module should run the remainder of the processing (calculation of different scores and decisions). Thus, making the right choice from among these thresholds will directly impact the efficiency of the approach retained as well as its global cost (CPU, memory, energy, etc.). We note that in some cases, τL could be different from τH.
1.4.2.2. Stage 2: classifying channels depending on the energy cost of the transmission models
Our aim is to minimize energy consumption by choosing the channels that use least energy. To do this, we classify the candidate channels depending on their energy score.
On each Ci channel, it is possible to transmit with the help of m modulations Mm. (MOD)im designates the transmission models possible on each channel Ci, where i represents the channel and m designates the modulation schema and the coding. Because the transmission power of the object has an impact on energy consumption, we suggest evaluating the transmission cost of the transmission models per available channel (MOD)im. This makes it possible to select the channel that most reduces energy consumption and so prolongs the battery life. In fact, the higher the transmission power, the better the signal-to-noise ratio, but a higher transmission power also means greater energy consumption (Krief 2012).
To select the most appropriate channel, the multicriteria decision-making module suggested should (1) calculate the minimum transmission power needed to transmit correctly on a given channel Ci with a modulation Mm, toward the other pair that could be an access point (M-RAN context) or another object (CRN context ), and (2) eliminate the transmission models (MOD)im necessitating a minimal transmission power greater than the device’s maximum transmission power.
To avoid rerunning the calculations at each execution of the selection algorithm, we suggest making calculations upstream and to store them within each object or on a remote device (Fog, Cloud, etc.). In fact, the minimum transmission power for each possible transmission model (channel, modulation) is calculated for different distances and stored on a Fog or Cloud device accessible to the objects.
The energy score is calculated according to the energy cost of a candidate channel and the battery life of the IoT device. For a given modulation and coding schema, the energy cost of a candidate channel is the average energy consumed per bit for the transmission on the channel. The energy cost of data transmission from the application j on this channel i (written Ci,j) is calculated depending on the band frequency, the uplink channel rate and the data flow from application (packet size and packet interarrival time). However, in the context of video streaming, other parameters can be considered to estimate this cost, such as video quality, the number of users and the quality of the channel (Zou et al. 2017).
The energy score SEnergy (i, j) is given by equation [1.6]:
where Ci,j represents the energy cost of transmission of data j from the application along the channel i and EB designates the object’s battery level.
Finally, the selected channels are classed depending on their energy scores. These are calculated according to the energy cost of the transmission models retained (MOD)im and according to the level of the battery of the IoT device.
1.4.2.3. Stage 3: calculation of the QoS score of the transmission models
In this stage, the multicriteria decision-making module should calculate the QoS score of the remaining candidate channels. The application’s QoS score SQoS(i, j) makes it possible to estimate the QoS offered to the application j when the transmission uses (MOD)im (channel i, modulation m).
SQoS(i, j) is based on the QoS parameters, that is the bandwidth, the loss rate, the delay and the jitter. The importance of these parameters depends on the application that is running. It is expressed by attributing weightings to different parameters.
The QoS SQoS(i, j) score is given by equation [1.7]:
where
- – Bi is the bandwidth available via the channel i;
- – bj is the minimal bandwidth required by the application j;
- – Ei is the loss rate via the channel i;
- – ej is the maximum loss rate authorized by the application j;
- – Di is the average delay in the channel i;
- – dj is the maximum delay supported by the application j;
- – Gi is the average jitter in the channel i;
- – gj is the maximum jitter supported by the application j;
- – vi is the maximum speed authorized on the channel i (vi depends on the network coverage);
- – V is the speed of the IoT object;
- – αB is the weight of the bandwidth;
- – αD is the weight of the delay;
- – αG is the weight of jitter;
- – αE is the weight of the loss rate;
- – xi⊗yj = 1 if xi > yi, otherwise 0;
- – Ai is the availability probability of the channel at the period t + 1.
The formula used to calculate the application’s QoS score j makes it possible to distinguish the different classes of service by adjusting the threshold values (bj, dj, gi, ej) and the weight criteria (αB, αD, αG, αE).
1.4.2.4. Stage 4: attribution of weight and calculating the final score
This last stage will be responsible for calculating the total score ST(i, j) for each candidate channel by weight the energy scores and the QoS. The weightings attributed to both scores will make it possible to select the IoT application domain (e.g. VANET and e-health).
This stage will therefore calculate the total score ST(i, j) for each candidate channel depending on the weight attributed to each aspect. The weight attribution for the criteria should reflect the importance of each decision criteria depending on the application context or, depending on the case, the preferences of the user (0 ≤ δ ≤ 1).
The total score ST(i, j) is given by equation [1.8]:
The selected channel will be the candidate with the highest total score. Finally, this information will be registered in the database for future use when a similar case occurs.
1.4.3. Performances evaluation
The multicriteria decision-making module that we suggest should be useful in different IoT contexts, particularly in M-RAN and CRN contexts. To illustrate the use of the retained approach in these two contexts and evaluate performances in each of them, we consider, in this section, two significant IoT use-cases: VANET and CR-VANET (Cognitive Radio VANET). In a VANET network, the On-Board Unit (OBU) is installed in the vehicle and includes a wireless transmitter receiver and various sensors, whereas a Road-Side Unit (RSU) is deployed in strategic locations along the route to facilitate communication between the vehicle and the infrastructure (V2I: Vehicle to Infrastructure). In V2I, information is exchanged between the vehicle and the RSU, or potentially a cellular network. VANET applications can be divided into three main categories (Javed et al. 2014): road safety (collision detection, cooperative driving, etc.), traffic management (route guidance, traffic light synchronization, etc.) and user infotainment (Web, streaming audio/video, etc.). Safety applications do not tolerate a transmission delay higher than 0.5 s (Javed et al. 2014), whereas traffic management applications are less demanding, with a tolerated latency between 0.1 and 1 s (Javed et al. 2014). As for infotainment applications, they generally accept greater latency, in order of 1–5 s. Nevertheless, some applications of this type, like multiplayer games, may require lower latencies of 0.1–1 s (Javed et al. 2014).
Based on the access networks detected or the channel available, and depending on the QoS requirements of the VANET applications, the suggested multicriteria decision-making module will select the best access network or radio channel available for a specific case. To do this, the utility function will calculate the scores of different candidate networks/radio channels to select the one that has the highest score.
Since we are considering combustion rather than electrical vehicles, we suppose that there is no energy constraint. Thus, the scores calculated from the different candidate channels are based only on the QoS parameters.
To evaluate the approach retained, we consider two representative types of transmission: multimedia emergency notification and infotainment applications, especially restaurant reservation. Table 1.1 summarizes the QoS requirements of these services.
Table 1.1. Some vehicular network services and the corresponding QoS parameters
| Type of service | Relevance of QoS parameters |
| Multimedia emergency notification | Latency (+++), bandwidth (+++) and packet loss (+) |
| Restaurant reservation | Latency (+), bandwidth (+) and packet loss (+) |
Table 1.2 shows the estimation of the weight of the data flow, the delay and the packet loss rate for services considered in the vehicle networks. These weightings are based on the importance of each parameter in Table 1.2.
Table 1.2. Weight estimation for services considered in vehicular networks
| Type of service | Weight estimation | ||
| Delay | BP | PLR | |
| Multimedia emergency notification | 0.43 | 0.43 | 0.14 |
| Restaurant reservation | 0.33 | 0.33 | 0.33 |
1.4.3.1. Scenario 1: network selection – the case of V2I communications
In this first scenario, we consider urban VANETs where the scene of the accident is captured by the first vehicle on the scene. Then, this vehicle will transmit the video to the rescue teams so that they can manage the emergency more effectively. In such a situation, the total transmission delay for a multimedia message is limited to a few seconds to ensure notification in real time (Javed et al. 2014). For this type of traffic (Table 1.1), the delay and the flow are considered dominant attributes.
As Figure 1.3 shows, the network architecture in a city is based on the vehicles with OBUs and infrastructures such as RSUs, WiFi access points and LTE eNB base stations for the 4G cell phone network. We suppose that each vehicle is equipped with three radio interfaces: 4G interface, WiFi interface and 802.11p interface.
Figure 1.3. Transmission of video from the scene of an accident in VANET networks
The simulation of this scenario, under the network simulator ns3, involves a WiFi access point and a LTE base station offering a theoretical flow of 11 Mbps and 24 Mbps, respectively. At the start, the number of active users is 4 for the WiFi cell and 10 for the LTE cell. The simulation parameters are listed in Table 1.3.
The effective data flow observed by the user of an LTE or WiFi network may be much lower than the theoretical flows stated and defined by the norms. The main factors influencing the effective flow are the number of active users sharing the bandwidth within a cell, the bandwidth frequency allocated to the network operator and the distance between the terminal and the relay antenna. These factors will also influence other QoS parameters such as latency and the average loss rate. For these reasons, the suggested multicriteria decision-making module will calculate the scores of the different networks detected depending on the estimated available bandwidth, the average delay measured and the packet rate loss measured.
Table 1.3. Simulation parameters for scenario 1
| Parameters | Values |
| Duration of simulation | 50 s |
| Traffic video | Send interval 1 ms Packet size 1,000 octets |
| Witness vehicle |
Speed: stationary Distance – AP WiFi: 20 m Distance – LTE BS: 60 m |
| Mobility model | Random Waypoint Model |
| Number of active users |
At t = 10 s, WiFi (4), LTE (10) At t = 20 s, WiFi (10), LTE (20) At t = 30 s, WiFi (20), LTE (10) At t = 40 s, WiFi (10), LTE (4) At t = 50 s, WiFi (10), LTE (10) |
| Theoretical flow for WiFi | 11 Mbps |
| Theoretical flow for LTE | 24 Mbps |
The number of active users may vary during the simulation (Table 1.3) as vehicles or pedestrians depart or arrive. The variation in this parameter influences the QoS parameters characterizing the candidate networks and, consequently, their scores. Figure 1.4 shows the effect of the dynamic network environment on optimal network selection for the service video. Depending on the configuration adopted for this scenario, the results of the simulation show that LTE is selected as an optimal network. In fact, when the LTE score increases because of the dynamic network environment, the optimal network passes from the WiFi to the LTE for the network selection decision. To ensure the stability of the decision and avoid the ping-pong effect, we wait for the following period to make the decision to be sure that the right decision is made.
Considering this example, if (1) the execution period of the suggested decision-making module is not adjusted automatically and (2) the transfer is made after the new trend is confirmed, the transfer will take place at 40 s. This is much later than the changeover time estimated at 23 s. This is why, as we said previously, it is very important to adjust the execution period of the suggested decision-making module automatically.
Figure 1.4. Score variation in VANET networks for transmission of a video message
1.4.3.2. Scenario 2: selecting the radio channel – the case of CR-VANET
Thanks to its cognitive capacity, the compatible intelligent radio vehicle can search for spectrum holes (white spaces) and take opportune decisions to transmit without interfering with the primary users. It can also make adaptations such as modifying the transmission power or the radio channels to meet QoS requirements for vehicle applications.
In this second scenario, we will consider a CR-VANET highway environment in the countryside and we will focus on infotainment applications, especially restaurant reservation. On the topology of CR-VANET highways, the speed of vehicles on the highways is significantly higher than in urban scenarios.
To evaluate performances, we consider the network architecture of a CR-VANET (Figure 1.5) where each vehicle is equipped with wireless communication interfaces (such as a 4G interface and a WiFi interface) and has intelligent radio capacities.
The simulations were made using ns3 paired with SUMO to generate real traffic mobility. The simulation parameters are listed in Table 1.4. For the same reasons cited in the previous scenario, the energy aspect will not be considered here.
Figure 1.5. Entertainment service on a highway in CR-VANET networks
Table 1.4. Simulation parameters for scenario 2
| Parameters | Values |
| Simulation duration | 50 s |
| Data traffic | Flow: 64 kbps |
| Vehicle mobility model | Car Following Model |
| Intelligent radio channels available |
TV white space band WiFi white space band LTE white space band |
For the vehicle behavior model, we have adopted the Car Following Model, which is based on the following parameters: vehicle acceleration, vehicle deceleration, vehicle length, maximum speed of the vehicle and driver imperfection. The values of these parameters are presented in Table 1.5.
Based on the spectral resources detected, the suggested multicriteria decision-making module will calculate the scores of different channels and select the one with the highest score.
Table 1.5. Simulation parameters for the SUMO vehicle
| Parameters | Values |
| Vehicle acceleration | 2.5 m/s2 |
| Vehicle deceleration | 4.6 m/s2 |
| Average vehicle length | 5 m |
| Maximum vehicle speed | 140 km/h |
| Driver imperfection | 0.5 |
Figure 1.6. Variation of the scores of intelligent radio channels for the entertainment service
Figure 1.6 shows the impact of the dynamic network environment (such as the available bandwidth, the channel’s availability probability and the speed of the vehicle) on selecting the channel best adapted to the infotainment service (restaurant reservation). In Figure 1.6, we can observe that with the high mobility of candidate vehicles on the highways, the channel corresponding to the WiFi bandwidth is not a good candidate. We also observe a degradation of the score of the radio channel corresponding to the bandwidth of the LTE and an increase in the score of the channel corresponding to the TV bandwidth. The evolution of the scores over time justifies the decision to carry out a spectrum handoff, from the suggested module to the TVWS channel. Before running the spectrum handoff, we wait for the choice of the best channel to be confirmed for the following period. This will enable us to avoid unhelpful transfers. And to avoid running the spectrum handoff very late, this wait time will be adjusted automatically. In this scenario, we observe that the choice of TVWS is fully adapted to this context, not only in terms of QoS but also from the perspective of mobility. Indeed, this spectral band is characterized by a long range that is best suited to scenarios with high mobility.
1.5. Conclusion
In this chapter, we tackled the question of decision-making for effective access to a radio network or a spectrum band in the IoT. An IoT object, having several interfaces and/or with cognitive capacities, can detect several access networks or radio communication channels. Thus, it may be led to choose the access network or radio communication channel that best meets the QoS constraints of the IoT application as well as its energy constraints. Selecting the most appropriate network or radio communication channel will allow the object to remain best connected. In this chapter, we focused on the functioning of the multicriteria decision-making module that we have suggested to tackle the problem of scalability. However, many approaches (Lounis et al. 2012; Gia et al. 2015; Guo et al. 2017; Firouzi et al. 2018; Shrestha et al. 2018; Khan and Lee 2019) have been proposed to solve the scalability issues in an IoT system. In our context, this question remains very important. This is why we plan to study it in future work.
1.6. References
Aijaz, A. and Aghvami, A.H. (2015). Cognitive machine-to-machine communications for Internet-of-Things: A protocol stack perspective. IEEE Internet of Things Journal, 2(2), 103–112.
Akyildiz, I.F., Lee, W.-Y., Vuran, M.C., Mohanty, S. (2006). NeXt generation/dynamic spectrum access/cognitive radio wireless networks: A survey. Computer Networks Journal, 50(13), 2127–2159.
Al-Fuqaha, A., Guizani, M., Mohammadi, M., Aledhari, M., Ayyash, M. (2015). Internet of Things: A survey on enabling technologies, protocols and applications. IEEE Communications Surveys & Tutorials, 17(4), 2347–2376.
Atzori, L., Iera, A., Morabito, G. (2010). The Internet of Things: A survey. Computer Networks, 54(15), 2787–2805.
Bhute, H.A., Karde, P.P., Thakare, V.M. (2014). Vertical handover decision strategies in heterogeneous wireless networks. Proceedings of International Conference on Recent Trends in Information, Telecommunication and Computing, ITC 2014. doi:02.ITC.2014.5.71.
Carneiro, P., Fortuna, P., Ricardo, M. (2009). Flow monitor: A network monitoring framework for the network simulator 3 (NS-3). Proceedings of the 4th International ICST Conference on Performance Evaluation Methodologies and Tools (VALUETOOLS ’09). doi:10.4108/ICST.VALUETOOLS2009.7493.
Firouzi, F., Farahani, B., Ibrahim, M., Chakrabarty, K. (2018). Keynote paper: From EDA to IoTeHealth: Promises, challenges, and solutions. In IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems. doi:10.1109/tcad.2018.2801227.
Gia, T.N., Rahmani, A., Westerlund, T., Liljeberg, P., Tenhunen, H. (2015). Fault tolerant and scalable IoT-based architecture for health monitoring. In IEEE Sensors Applications Symposium (SAS). IEEE, Zadar.
Grace, D., Chen, J., Jiang, T., Mitchell, P.D. (2009). Using cognitive radio to deliver “green” communications. In 2009 4th International Conference on Cognitive Radio Oriented Wireless Networks and Communications. IEEE, Hanover.
Guo, H., Ren, J., Zhang, D., Zhang, Y., Hu, J. (2017). A scalable and manageable IoT architecture based on transparent computing. Journal of Parallel and Distributed Computing, 118(1) 5–13. doi:10.1016/j.jpdc.2017.07.003.
Haykin, S. (2005). Cognitive radio: Brain-empowered wireless communications. IEEE Journal on Selected Areas in Communications, 23(2), 201–220.
Iqbal, J., Khan, M., Afaq, M., Ali, A. (2019). Performance analysis of vertical handover techniques based on IEEE 802.21: Media independent handover standard. Transactions on Emerging Telecommunications Technologies. doi:10.1002/ett.3695.
Javed, M.A., Ngo, D.T., Khan, J.Y. (2014). Multimedia communication for emergency services in cooperative vehicular ad hoc networks. In Multimedia over Cognitive Radio Networks, Hu, F., Kumar, S. (eds). CRC Press, Boca Raton.
Kassar, M., Kervella, B., Pujolle, G. (2008). An overview of vertical handover decision strategies in heterogeneous wireless networks. Journal of Computer Communications, 37(10), 2607–2620.
Khan, U.A. and Lee, S.S. (2019). Multi-layer problems and solutions in VANETs: A review. Electronics 2019, 8, 204.
Khan, A.A., Rehmani, M.H., Rachedi, A. (2016). When cognitive radio meets the Internet of Things? In 2016 International Wireless Communications and Mobile Computing Conference (IWCMC). IEEE, Paphos.
Koushik, A.M., Hu, F., Kumar, S. (2018). Intelligent spectrum management based on transfer actor-critic learning for rateless transmissions in cognitive radio networks. IEEE Transactions on Mobile Computing, 17(5), 1204–1215.
Krief, F. (ed.) (2012). Green Networking. ISTE Ltd, London, Wiley, New York.
Kumar, K., Prakash, A., Tripathi, R. (2016). Spectrum handoff in cognitive radio networks: A classification and comprehensive survey. Journal of Network and Computer Applications, 61, 161–188.
Kumar, K., Prakash, A., Tripathi, R. (2017). A spectrum handoff scheme for optimal network selection in NEMO based cognitive radio vehicular networks. Wireless Communications and Mobile Computing, 1–16.
Lounis, A., Hadjidj, A., Bouabdallah, A., Challal, Y. (2012). Secure and scalable cloud-based architecture for e-health wireless sensor networks. In International Conference on Computer Communication Networks (ICCCN). IEEE, Munich.
Perera, C., Liu, C.H., Jayawardena, S., Chen, M. (2014). A survey on Internet of Things from industrial market perspective. IEEE Access, 2, 1660–1679.
Perera, C., Liu, C.H., Jayawardena, S. (2015). The emerging Internet of Things market-place from an industrial perspective: A survey. IEEE Transactions on Emerging Topics in Computing, 3(4), 585–598.
Shah, M.A., Zhang, S., Maple, C. (2013). Cognitive radio networks for Internet of Things: Applications, challenges and future. In 19th International Conference on Automation & Computing. IEEE, London.
Shaikh, F.K., Zeadally, S., Exposito, E. (2017). Enabling technologies for green Internet of Things. IEEE Systems Journal, 11(2), 983–994.
Shrestha, R., Bajracharya, R., Nam, S.Y. (2018). Challenges of future VANET and cloud-based approaches. Wireless Communications and Mobile Computing. doi:10.1155/2018/5603518.
Singh, K.D., Rawat, P., Bonnin, J. (2014). Cognitive radio for vehicular ad hoc networks (CR-VANETs): Approaches and challenges. Journal on Wireless Communications and Networking, 49(2014). doi:10.1186/1687-1499-2014-49.
Song, Y. and Xie, J. (2012). ProSpect: A proactive spectrum handoff framework for cognitive radio ad hoc networks without common control channel. IEEE Transactions on Mobile Computing, 11(7), 1127–1139.
Tarek, A., Benslimane, M., Darwish, M. (2020). Survey on spectrum sharing/allocation for cognitive radio networks Internet of Things. Egyptian Informatics Journal. doi:10.1016/j.eij.2020.02.003.
Tuysuz, M.F. and Trestian, R. (2017). Energy-efficient vertical handover parameters, classification and solutions over wireless heterogeneous networks: A comprehensive survey. Wireless Personal Communications, 97, 1155.
Vlacheas, P., Giaffreda, R., Stavroulaki, V., Kelaidonis, D., Foteinos, V., Poulios, G., Demestichas, P., Somov, A., Biswas, A.R., Moessner, K. (2013). Enabling smart cities through a cognitive management framework for the Internet of Things. IEEE Communications Magazine, 51(6), 102–111.
Wang, L.-C. and Wang, C.-W. (2008). Spectrum handoff for cognitive radio networks: Reactive-sensing or proactive-sensing? In Proceedings of IEEE International in Performance, Computing and Communications Conference (IPCCC 2008), Austin, TX, 343–348. 10.1109/PCCC.2008.4745128.
Wang, Y.-H., Huang, G.-R., Tung, Y.-C. (2014). A handover prediction mechanism based on LTE-A UE history information. In International Conference on Computer, Information and Telecommunication Systems (CITS). IEEE, Jeju.
Wu, Q., Ding, G., Xu, Y., Feng, S., Du, Z., Wang, J., Long, K. (2014a). Cognitive Internet of Things: A new paradigm beyond connection. IEEE Internet of Things Journal, 1(2), 129–143. doi:10.1109/JIOT.2014.2311513.
Wu, Y., Hu, F., Kumar, S., Zhu, Y., Talari, A., Rahnavard, N., Matyjas, J.D. (2014b). A learning-based QoE-driven spectrum handoff scheme for multimedia transmissions over cognitive radio networks. IEEE Journal on Selected Areas in Communications, 32(11), 2134–2148.
Xenakis, D., Passas, N., Di Gregorio, L., Verikoukis, C. (2011). A context aware vertical handover framework towards energy-efficiency. In Proceedings of the IEEE Vehicular Technology Conference (VTC).
Xiao, K. and Li, C. (2018). Vertical handoff decision algorithm for heterogeneous wireless networks based on entropy and improved TOPSIS. In 2018 IEEE 18th International Conference on Communication Technology (ICCT). IEEE, Chongqing.
Xu, L.D., He, W., Li, S. (2014). Internet of Things in industries: A survey. IEEE Transactions on Industrial Informatics, 10(4), 2233–2243.
Zekri, M., Jouaber, B., Zeghlache, D. (2012). A review on mobility management and vertical handover solutions over heterogeneous wireless networks. Computer Communications, 35, 2055–2068.
Zhu, C., Leung, V.C., Shu, L., Ngai, E.C.H. (2015). Green IoT for smart world. Special Section on Challenges for Smart Worlds. doi:10.1109/ACCESS.2015.2497312.
Zou, L., Javed, A., Muntean, G. (2017). Smart mobile device power consumption measurement for video streaming in wireless environments: WiFi vs. LTE. In 2017 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB). IEEE, Cagliari.